A friend of mine recently started his Social Security benefits at age 66, justifying his decision because interest rates have risen recently from virtually zero to the four percent range. That …
I am told I have a knack for identifying companies making waves in specific technology areas. One such company is Kyndryl, marking its second anniversary of separation from IBM. I believe …
Banks Plan Payment Wallet to Compete With PayPal, Apple Pay Big banks are teaming up to launch a digital wallet that people can use to shop online. Wells Fargo, Bank of …
By now you’re well aware of the requirements to protect customer data against loss. You know that you can’t allow personally identifiable information to be exposed. You also know that the …
By now you’re well aware of the requirements to protect customer data against loss. You know that you can’t allow personally identifiable information to be exposed. You also know that the costs of doing so are significant – you or your company can find yourselves under indictment, you can face lawsuits and the damage to your company’s reputation will be serious to the point that you might not recover.
So why do a third of businesses still use real data when testing their company’s software or production environments? That’s the question that was approached in a study Tonic, which bills itself as “The Fake Data Company.”
The study found that 29 percent of the companies responding used unprotected production data (this is real customer or personnel data) in testing environments or when troubleshooting their company’s software. The study also found that 45 percent of respondents faced a major data breach due at least in part insecure data handling. The cost of those data breaches included delays in the company’s ability to function, reported by 88 percent of respondents.
And unfortunately, the problem is growing. “With the rising popularity of remote work, increased data breaches and leaks are inevitable. Whether it be stolen work laptops, using real customer data in testing, or the incorrect cloud migration of data, there are a myriad of reasons why a breach can happen,” said Ian Coe, CEO of Tonic.ai. “What’s important to understand is that while a breach may happen, there are ways to safeguard the data to ensure companies are not putting their customers’ information at risk.”
Fake Data
One solution to this risk of a data breach is to not use real data in the first place. If all of the data you use for troubleshooting and testing is fake, then a data breach in those areas won’t matter. For this you need fake data. Fake data appears to your data systems to be real, it can be processed as if it were real, but it’s not real.
Yoou can use fake data to protect the real thing
getty
The idea of using fake data to protect against a breach that otherwise would reveal real protected data has probably occurred to most IT managers, but there’s a problem – how to get enough fake data to provide a meaningful test. It’s one thing to turn a couple hundred data entries into a fake representation of your data manually, but beyond that it’s too much manual work to be economically useful.
Tonic.ai solves this problem by being able to produce vast quantities of fake data that’s customized to look just like your real data, except for being fake.
“Fake data is designed to respect and protect the privacy of real customer data. Companies should be responsible about their data governance and being compliant when it comes to things like personal identifiable information,” said Coe.
While using fake data can’t do anything about a company’s insecure testing practices or the lack of protection against data breaches, it can at least eliminate the chance that your test will reveal protected data, which in turn help keep you out of the courts or jail.
New mental health challenges are impacting our organizations now daily – And it is not easing up any time soon. Mental health since COVID has been deteriorating by 25% from Pre-COVID …
New mental health challenges are impacting our organizations now daily – And it is not easing up any time soon. Mental health since COVID has been deteriorating by 25% from Pre-COVID statistics. I was on a call yesterday with a former CEO of a major securities firm, and he simply said, everyone I know is depressed or have children feeling sad.
When you dig harder into the research as I have pointed out in prior blogs, we have 20-30 % of people worldwide waking up and starting their day – either sad or angry.
So, what are our board directors or C-levels doing to increase more focus on health and wellness of the organizations they are governing.
Dr. Robbie Babins, CEO of a Calgary Counselling Center has stated: I believe the pandemic is transformational in terms of the more open conversations we are having and the ways in which our communities and workplaces are approaching mental health and taking it action to make it better.
This being said, are we doing enough?
I don’t think so.
National Data released by Statistics Canada in August of this year, showed that the increase of depression disorders was two times higher than Pre-COVID, and that the percentage of adults with self reported depression has also doubled.
Wow and the USA is not fairing better either.
I think we can assume there are far greater numbers that what these numbers are revealing, as many employees are not comfortable reporting or seeking support which is an entirely other issue of educational awareness.
With the real estate shifts to more hybrid work, many companies are wrestling with inconsistent policies or outdated leadership behaviours that think if I can see them I know they are working for me. This is a going out of business mindset.
With the challenges in recruiting talent and keeping talent, here are some questions that board directors can ask of their CEOs or board committees:
1.) Does your board have a clearly defined strategy supporting health and wellness with clearly defined measurement systems?
2.) Does your board meeting always have a health and wellness report and how is your company compared to other organizations in your industry or competitive landscape?
3.) What innovative mental health and wellness tools are you purchasing to help communicate daily to your talent how much you care.
Conclusion:
Bringing mental health out of the shadows and into mainstream management with effective and frequent leadership communication is a step in the right direction. What we do know is the solution is not an HR employee satisfaction survey – today real time communication is needed – nothing less will suffice.
So far the statistics are continuing to show more risks and more sadness and unhappiness in 2024, so every board director and C level has to reflect on how you are going to show more love to your employees, as the Great Resignation Tidal wave that is hitting our corporation doorsteps is persisting.
Love to hear your thoughts – so reach out on my Linked in Channel if you have a story about a best practice – I am working on a new book on Happiness Counts and its Big Business, so primary research is important.
How do you pivot your career into completely different industries? originally appeared on Quora: the place to gain and share knowledge, empowering people to learn from others and better understand the …
How do you pivot your career into completely different industries? originally appeared on Quora: the place to gain and share knowledge, empowering people to learn from others and better understand the world.
Answer by David Garcia, President of New Stand, on Quora:
For many, the last several years have created not only major transformations in the workplace but in how we think about our relationship with work. Millions of people are using this as an opportunity to change the course of their careers.
If you’re hitting the job market, either by choice or by force, pivoting is one of the most rewarding and challenging moments of your career. Rarely does life give you the chance to recreate who you are and give the opportunity to put to work all of the learning you’ve accumulated without any of the baggage.
Here’s a quick explainer on how to make one of the biggest moves of your career.
Define what WINNING means to you.
Set your north star by asking yourself why you want to make the move and what outcomes you want to achieve. Whether it’s creative fulfillment, financial gain, or working on something with meaning, set your north star and don’t accept a job if you don’t win.
Bet on the right horse.
Pick high growth industries and do your homework before you place a bet — if you’re pivoting into emerging tech (such as, crypto, drones, and VR) these are radically different markets and you have to understand the advantages and disadvantages of the markets and players BEFORE you make a move.
No one wants to hire a college grad.
Self-educate because the biggest hurdle you’ll have to cross is “Do I think this candidate can do the job?” Demonstrating understanding of core concepts and second/third order effects will go a long way in getting sponsors to refer you, recruiters to screen you, and hiring managers to select you over other candidates.
Double down on betting on yourself.
If getting into the company or industry matters more than immediate financial gain or scope, don’t be afraid to accept a lateral move or take a step back — you’re already betting on yourself. Breaking in through the proverbial mailroom is still better than not breaking in at all and once you’re in you can show and prove your way to the top.
The internet is not your friend.
Your candidacy runs counter to all pattern recognition at play in the system — you will be automatically disqualified by applicant tracking systems and deprioritized by algorithms. Leverage your network to get intros and meet people. If you don’t have a network to activate, cold email and call your way into the recruitment process.
Always be closing.
Every interaction is part of the interview process, sell yourself in every moment until they start selling you. Key to getting over the goal line is addressing your perceived weaknesses. Take the 8-Mile approach and be the first to raise potential concerns: “I know I don’t have the healthcare background that you’re usually looking for, but let me tell you why I’m still the best candidate…”
You made it happen, don’t mess it up.
The first 90 days are critical in demonstrating to the hiring manager and company they made the right decision in selecting you, don’t mess it up.
This question originally appeared on Quora – the place to gain and share knowledge, empowering people to learn from others and better understand the world.
I have written a few blogs recently on the digital literacy gaps in our businesses and society at large, as we continue to see Artificial Intelligence solutions in all sectors hitting …
I have written a few blogs recently on the digital literacy gaps in our businesses and society at large, as we continue to see Artificial Intelligence solutions in all sectors hitting our global markets.
Digital literacy has become indispensable for every global citizen, whether to communicate, find employment, receive comprehensive education, or socialize.
Acquiring the right set of digital skills is critical for learning and workforce readiness but also vital to foster more open, inclusive, and secure ways of connecting with one another.
Digital Literacy (DL) is a hot topic in the boardrooms as the cry out Big Q is: how literate are our board directors in terms of their digital skills and AI acumen? That’s one key point, the other is what is Digital Literacy, as it’s still a term in constant evolution and likely will be until the end of time.
Getting a handle on an agreed upon definition that a board or an organization can rally around is an important first step. So here are a few DL points of view, from different international sources, including: UNESCO, the European Union, Ireland and Canada, etc.
UNSECO states there is no one set of agreed definitions for digital literacy which is an interesting point. The Irish National Forum for the Enhancement of Teaching and Learning in Higher Education has identified over 100 models and frameworks of digital skills, literacies or competencies. See Ireland’s All Aboard Digital Literacy Framework for Higher Education here.
So this may sound like a messy question to answer.
An excellent digital literacy framework is the DigComp developed by the Joint Research Centre (JRC) of the European Commission.DigComp 2.1 identifies five key components of digital competence: information and data literacy, communication and collaboration, digital content creation, safety, and problem solving, with also 21 related competences and eight proficiency levels.
Even British Columbia in Canada has a framework for Digital Literacy for K-12. This framework, and the competencies in it are categorized by grade level. This still applies well to higher education (simply remove the grade levels – the college student should learn all of these competencies). The BC framework defines: “Digital Literacy is the interest, attitude and ability of individuals to use digital technology and communication tools appropriately to access, manage, integrate, analyze and evaluate information, construct new knowledge, and create and communicate with others.”
An often cited framework for information literacy for higher education is from the Association of Colleges and Research Libraries. See Framework: ACRL. Each of these links below helps to inform the documented framework.
It is useful to review the responsibilities of a Board of Directors, especially those of a publicly traded company. Boards are primarily responsible for looking after the interests of shareholders. They accomplish this by providing oversight in three key areas:
Strategic direction and advice: As Technology is underpinning all business strategy, it is impossible to provide strategic direction without looking at the technology enablements.
Financial oversight: No CTO or CIO will debate how expensive IT and complex SCM solutions are to embed in an organization, and the oversight on these high risk investments require astute technology and business leaders working as partners, and digital literacy is critical for effective communication.
Risk management: The risks of technology investments are tightly integrated with business strategy and financial management practices.
As Data, AI and CyberSecurity are top of mind digital literacy topics, the imperative to ensure all board directors are recruited with a strong digital literacy foundation is key as the world continues to modernize faster and faster with collective data intelligence.
Where is your board at in digital literacy proficiency? How well prepared is your organization?
Picture this: You are age 72 or older and you are still working and participating in your company’s 401(k) plan. Did you know that you may be able to delay required …
The rules are set out in Internal Revenue Code Section 401(a)(9)(C). The 401(k) must be your current employer’s plan, not a former employer’s plan; you cannot own more than 5% of your current employer; and the plan has to allow the delay. (As IRS Publication 575, “Pension and Annuity Income,” points out, “your plan may require you to begin to receive distributions by April 1 of the year that follows the year in which you reach age 72 even if you haven’t retired.”)
This “still-working” exception for 401(k)s does not apply to IRAs.
When Is The First RMD For A 401(k) Rollover?
Since each person’s tax situation is unique, you’ll need to consult with your tax adviser for answers that apply to you.
These reader situations can help illustrate how the rules work. I reached out to Alex Kotwal, a CPA and tax manager for Marcum LLP, a national accounting and advisory firm, for some general insights.
In one case, 75-year-old “Wayne” hasn’t retired yet; in fact, he is still contributing to his 401(k) at work. Even though he is over age 72, Wayne is not taking RMDs from the 401(k) – he doesn’t need to under the still-working exception to RMDs.
Wayne wants to know if he can do an in-service rollover from his 401(k) to an IRA, which is permitted by his plan. He wants to time the rollover before the end of the year, while he is still working. Wayne’s question: Will the rollover to the IRA trigger an RMD for 2022?
According to Kotwal, here’s the general rule: “An IRA’s RMD is based on the value of the account on Dec. 31st of the previous year. Since the IRA didn’t exist as of 12/31/2021, there would not be an RMD requirement for 2022. The first RMD would begin in 2023.”
What Happens To Your 401(k) When Your Company Is Sold?
In another situation, “John” is over age 75. The company where he was working, and where he had a substantial 401(k), was purchased by another company. All of the employees of the old company were terminated after the sale.
However, John was hired by the new company, and he is still working. The new company has its own 401(k), to which John can contribute.
However, John is looking to minimize any new RMDs. His existing 401(k) account is with a multinational financial services company, which informed him that he has options due to the company being sold: He can cash out his 401(k), roll it into an IRA, roll it over into his new employer’s 401(k), or leave it with the financial services company.
John would like to leave the 401(k) with the financial services company because he likes the conservative option. If he does that, he is wondering if and when he will need to take an RMD.
When Does The RMD Begin?
Kotwal points out that if John leaves the 401(k) with the financial services company, then he would be required to take RMDs from the account, as “the still-working exception only applies to the plan of the company which he is currently working for.” The required beginning date (RBD) for 401(k) RMDs is April 1of the year after John’s employment ended with the previous employer — in this case, 2023.
The solution for John is to roll over the old 401(k) into the new company’s 401(k). That is, as Kotwal noted, John “could continue to utilize the still-working exception by rolling over the 401(k) into the new employer’s plan.”
Caution
For situations like these, it is important to know the rules related to your 401(k) plan. Ask your plan sponsor or human resources department for details if you do not have them. And, one more time, seek the advice of your own CPA.
Can The 5% Ownership Rule For RMDs Be Reversed?
Another reader presented this situation: A 401(k) participant owned 100% of his company until 2021, when he sold all of his stock. He continued working for the buyer. He had been taking RMDs since age 70 1/2 because he did not qualify for the still-employed exemption (due to his ownership).
Now that he is working for the new company but is no longer a shareholder, can he take advantage of the still-working exemption? Or does he have a 2022 RMD?
Sorry, once a shareholder hits his RBD, there is no turning back. Experts at Fidelity Investments explained it this way: “A 5% owner is determined in the year in which the participant would have reached his/her required beginning date, either age 70 ½ (if he/she had attained age 70 ½ before Jan. 1, 2020) or 72 (if he/she had not attained age 70 ½ until after Dec. 31, 2019). If the participant is a 5% owner in the year of his/her required beginning date, then he/she will always be considered a 5% owner even if he/she thereafter sells his/her ownership in the company and will be required to continue receiving at least annual required minimum distributions.” Fidelity stresses, as I do, that the participant needs to confirm this answer with his or her tax adviser.
Questions?
To keep up with topics that I cover, be sure to follow me on the forbes.com site (and if you would like to subscribe, check out the red box at the top right). Write to me at [email protected]. Include your city and state, and mention that you are a forbes.com reader. While all questions cannot be answered, each email is read and reviewed and can lead to discussion in a future post.
Discussing money with loved ones is asking for a loaded conversation. Money is basically about relationships and it is nearly impossible to be in love without money playing a role. A …
Discussing money with loved ones is asking for a loaded conversation. Money is basically about relationships and it is nearly impossible to be in love without money playing a role. A new book by Stanford professor emeritus, Myra Strober, who is an early feminist economist, and ‘social innovation leader’ Abby Davisson called Money and Love: An Intelligent Roadmap for Life’s Big Decisions explores how to navigate the crossroads of finances and human bonds.
I had the pleasure of talking to Professor Strober this week on Zoom. I sat in my teensy New York kitchen and she in a study with the Palo-Alto sun streaming on colorful couch pillows.
I asked her, “In a world full of financial planners, more schooled in financial products and tax law than economists, what can an economist offer about financial planning a financial planner can’t?”
Polite Professor Strober nearly rolled her eyes and I thought I blew the interview. “I wrote the book because I am an educator.” Her own life experience and those of her thousands of students made her realize people made life decisions without understanding how budgeting and financial plans can be derailed by clueless relationships and the best relationships can be derailed by financial miscommunication and ignorance.
Indeed, Davisson and Strober’s book spans the life course: from the time you move into together, choose careers, and decide to be parents to the moment you discuss your mother’s retirement plans and move your spouse to memory care.
The part of Money and Love about cross generational conversations was uncomfortable, like a good teeth cleaning is uncomfortable. The necessary questions to ask your aging parents seem a little cringy. Strober acknowledges the costs of these seemingly prying and bossy questions but lifts-up the benefits. She steered me in the conversation about HOW to have the conversations. The advice is “slowly, lightly, and mindful of your own biases and concerns.”
Since we are interconnected, the difficult conversations are necessary. In the U.S., thirty percent of adults with a living parent or parents over the age of 65 says their parent or parents needs help handling their affairs or caring for themselves. For families who are not rich, the adult child is likely to provide the time and finances. This fact begged the question, are daughters and daughter in law more likely to forego paid work than a male relative? Strober replied, “well of course, women always do more work; but sons and sons in law are doing more and more.” And who ever does the work is paying for it indirectly.
The adult kids who help their parents spend about an average of 77 hours per month – to do so. Time doing elder care has an opportunity cost in the United States amounting to $522 billion annually.
Strober and Davisson write, “If you think it likely that you may be called upon to help your parents financially at some point, it will be helpful if you can get relevant information from them in advance. If they resist sharing details of their finances, this may be a tough assignment. Hopefully, you can prevail by explaining that you will be far more successful in providing whatever help they may need if you have some advance “warning” to do some planning yourself.”
They write that adult children should ask where their parents’ computer passwords are, their wills and the names of their accountant and attorney (if they have them). I wondered how you can have that conversation and not come across as noisy and bossy. Maybe just ask where they record their phone’s password so if they are hit by a bus you can find it. But it occurs to me good relationships involve reciprocity. Just because an adult child is younger doesn’t’ mean fatal accidents and illnesses can’t happen.
I asked her what the top five areas adult children are should discuss with their parents. I got three. First discuss your parent’s expectations about retirement; ask about their advance directives and wills and know where their passwords are.
I asked Professor Strober what the most important message of the book about difficult conversations was. “Tread lightly,” she said.
We both winced and almost said at the same time. Make sure you do the same for them. Be reciprocal. The idea here is that unless you organize your passwords, make your will, plot out some retirement planning you can’t badger your parents about it. Don’t ask your parents to do anything you haven’t done.
Money and Love is readable for anyone but can be a lively primer for clinical psychologists and psychiatrist whose love-torn patients need financial tips. It is also a good primer for financial planners whose clients need relationship tips.
Working from home (WFH) has taken a big jump after the Covid-19 pandemic, and some very smart people think it’s here to stay (as I reviewed in my last blog), because …
Working from home (WFH) has taken a big jump after the Covid-19 pandemic, and some very smart people think it’s here to stay (as I reviewed in my last blog), because many highly educated workers like it. But others aren’t so sure whether employers will continue supporting it, especially if the economy softens and labor markets get tighter.
Some employers don’t like WFH. In June, Elon Musk told Tesla employees they had to “spend a minimum of 40 hours in the office per week,” and not in “some remote pseudo-office.” If they didn’t come in, Musk and Tesla “will assume you have resigned.” And Musk now has cancelled Twitter’s WFH option, saying workers needed to be in at least 40 hours per week. Any exception has to be approved personally by Musk.
Musk isn’t alone. JP Morgan CEO Jamie Dimon has dismissed remote work and Zoom meetings as “management by Hollywood Squares,” seeing it feeding “a working environment that’s less honest and more prone to procrastination.” In May, Goldman Sachs CEO David Solomon said remote work was “an aberration” that didn’t fit Goldman’s “innovative, collaborative apprenticeship culture.” In October, Solomon said 65% of employees are back, compared to a pre-pandemic level of 75%.
But other firms report moving extensively to remote work, especially tech companies like Facebook, Amazon AMZN , and Microsoft MSFT . Human resources consultants say firms that won’t allow WFH or some flexibility risk “missing out on 50% to 70% of candidates,” especially tech and other highly skilled managers and experts.
A recent conference at the Boston Fed had a first-rate discussion of these issues. Economists Matthew Kahn and Steven Davis argued remote work is here to stay, with firms learning to adjust and even finding added value from new work arrangements.
But the conference also heard from the Wharton School’s Peter Cappelli, one of our best thinkers on work, management, and organizations. Kahn and Davis focused more on how workers prefer WFH, but Cappelli asked the other side of the question—“Is ‘Permanent Remote’ Good For Employers?” (The presentations can be found here, and a video of the full panel discussion here, starting at the two hour mark.)
Cappelli reviewed the experience with remote work prior to the pandemic. While “life satisfaction” increased for WFH employees, “work-related and career outcomes are worse on every dimension examined,” with increased communication problems at the firm and “more work for supervisors.”
Cappelli put the issues raised by Goldman’s CEO into a larger context. He noted that with “permanent remote” work, “a lot of things are harder to do—collaboration, innovation, maintaining culture, employee engagement, etc.” “Hybrid” work, where employees are in the office some of the time, isn’t an automatic solution. Firms still have to maintain office space, they get less social interaction and group creativity, and face increased scheduling and IT problems and more cumbersome project management.
Employers face a number of workforce challenges with extensive WFH. How are new employees onboarded and brought into the culture? What happens to on-the-job learning, a major source of how employees learn new skills, especially at a specific company? (Think about a new IT worker getting help from an experienced colleague at the next desk when facing a technical problem that needs solved right away, versus trying to connect remotely.)
For me, one major takeaway from the Boston Fed panel is how WFH is a potential locus of conflict between workers and employers. The initial health concerns in the pandemic and the tight labor market since gave workers—especially higher educated ones with essential skills—a lot of leverage with employers.
But that power balance may be shifting. If the Federal Reserve gets the recession it seems to want, then the labor market power balance will tilt more to employers. For example, some tech firms allowed significant working from home. As they lay workers off, and new job openings fall, employers will have more say over where and how the remaining staff will work.
There’s no question working from home has increased since the pandemic, and some of that increase will be permanent. But it also is tied to the ongoing and shifting power balances between workers and employers. The next few years, especially if we get a recession, will tell us how widespread and persistent this labor market change really is.
This week NeurIPS 2022 happens, one of the most important events of the year where the latest in AI research is presented. The topics covered by the keynotes are a great …
This week NeurIPS 2022 happens, one of the most important events of the year where the latest in AI research is presented. The topics covered by the keynotes are a great sample of the key preoccupations of the AI research community today: From whether large language models are sentient to the forward-looking alternative to backpropagation.
Sentient or not, MineDojo got Nvidia’s researchers the NeurIPS 2022 Outstanding Datasets and Benchmarks Paper Award, demonstrating their belief that large language models in the future will be “embodied agents that proactively take actions, endlessly explore the world, and continuously self-improve.” The current version of this AI agent learned Minecraft’s flexible gameplay using a massive online database of more than 7,000 wiki pages, millions of Reddit threads and 300,000 hours of recorded gameplay. (Here’s a Twitter thread summarizing all 15 Outstanding papers, written by Nvidia’s Jim Fan).
More than 59 zettabytes (59 trillion gigabytes) of data will be created, captured, copied, and consumed in the world this year, says IDC. The explosion of online, easily-available data leads to a constant stream of new applications of large language models.
Meta’s AI agent Cicero integrates a language model with planning and reinforcement learning algorithms by inferring players’ beliefs and intentions from its conversations and generating dialogue in pursuit of its plans. Across 40 games of an anonymous online Diplomacy league, Cicero achieved more than double the average score of the human players and ranked in the top 10% of participants who played more than one game.
Meta’s Yann LeCun proudly announced galactica.ai, “A Large Language Model trained on scientific papers. Type a text and galactica.ai will generate a paper with relevant references, formulas, and everything. Amazing work by @MetaAI.” To which Gary Marcus replied: “Pitch perfect and utterly bogus imitations of science and math, presented as the real thing. Is this really what AI has come to, automatically mixing reality with bullshit so finely we can no longer recognize the difference?” After three days of similar criticisms, Meta took down the public demo.
Indeed, there’s more work to be done, says Andrew Ng: “Some engineers (including the Galactica’s team) have proposed that LLMs could be an alternative to search engines. For example, instead of using search to find out the distance to the Moon, why not pose the question as a prompt to a language model and let it answer? Unfortunately, the maximum-likelihood objective is not well aligned with the goal of providing factually accurate information. To make LLMs better at conveying facts, research remains to be done on alternative training objectives or, more likely, model architectures that optimize for factual accuracy rather than likelihood.”
Attempting to organize the explosion of large language models (or foundation models) and to make transparent what is known about them, Stanford announced a new benchmark project covering 30 prominent language models across a wide range of scenarios and for a broad range of metrics. “We need to know what this technology can and can’t do, what risks it poses, so that we can both have a deeper scientific understanding and a more comprehensive account of its societal impact,” says the Center for Research on Foundation Models at the Stanford Institute for Human-Centered AI. Of their 5 key findings, I liked most that “human evaluation [is] essential in some cases.”
While NeurIPS showcases the recent work of large centralized labs owned by Big Tech, there has been a dramatic increase this year in the frequency of open-source releases of the latest AI breakthroughs by small independent labs. This is one of the major conclusions of the 2022 edition of the State of AI Report from Nathan Benaich and Ian Hogarth. They highlight as prime examples “small, previously unknown labs like Stability.ai and Midjourney,” demonstrating a “fundamental shift in the previously accepted AI research dynamic that larger labs with the most resources, data, and talent would continually produce breakthrough research.”
Another conclusion of the report is that China continues to lead the world in AI research (measured by the number of papers published), and “significantly leading” in areas with implications for security and geopolitics.
The State of AI report defines AI Safety as “a field that studies and attempts to mitigate the catastrophic risks which future AI could pose to humanity” and it reports that an estimated 300 safety researchers are working at large AI labs, compared to under 100 in last year’s report. But, says Andrew Ng, “without a broad international agreement on definitions of dangerous technology and how it should be controlled, people in other parts of the world will be free to ignore them. Considering the challenges, perhaps the best we can do is to work proactively and continually to identify potential misuses and ways to thwart them.”
Possibly not rising to the level of “catastrophic risks,” AI bias is very much on the mind of the AI community as one negative side effect of AI programs in general and large language models in particular.
The large language models that enable the mass deployment of text-to-image generation “amplify dangerous and complex stereotypes” and are a “cause for serious concern.” Experimenting with Stability.ai’s Stable Diffusion, the researchers found that the ”emotional person” was always a woman, “a terrorist” was always Middle Eastern in appearance, and “a thug” was always a Black man. The team at the HuggingFace blog has been working to improve their AI bias evaluation tool to include more nuances of discrimination. Jack Clark says that eventually, there will be as many approaches to evaluating AI bias as there are ideologies and countries. Tim O’Reilly says we can fix AI bias by first fixing ourselves.
On a more positive note, the application of AI models and algorithms to practical problems continues apace. AI-based retinal imaging and diagnostics startup AEYE Health received clearance from the FDA to market its diagnostic screening system for diabetic retinopathy. Using large language models, Seek AI automates the repetitive work that data professionals perform, improving their productivity.
This has been a random selection of recent news from the world of AI, defined as computer-assisted learning from data.
Too often, trust in an AI system is equated to explainability. What does the algorithm do? How does it work? Can I explain its outputs reasonably simply? If so, we tend …
Too often, trust in an AI system is equated to explainability. What does the algorithm do? How does it work? Can I explain its outputs reasonably simply? If so, we tend to consider an AI system trustworthy. If one is building a recommendation engine that suggests to a salesperson a particular action as next best, one must understand how the algorithm is reaching its conclusion. What data and information is it learning from? And how can we effectively communicate this to the salesperson so that they trust the answer and act on it? Similarly, if one is predicting which customers are likely to churn (or demonstrate increased propensity to buy), then one must consider why the model is suggesting this and how to prevent churn.
Explainability is a discipline unto itself, with leading academics and practitioners working to decode the inner workings of ever more complex algorithms. Along with explainability, another rising field is the study of AI bias. Bias usually stems from the data sets used to train AI algorithms. Are these sets truly representative, or do they disadvantage a subpopulation via underrepresentation? Do they propagate poor decision-making by training algorithms on suboptimal behavior from the past? Bias can rightfully erode trust in AI systems. And it need not only arise in complex systems — even simple AI applications can demonstrate bias due to bad data or other flaws.
But sometimes trust has nothing to do with the algorithms or the data – it exists outside of the AI system. Let me illustrate.
I recently visited Haiyi Zhu, the Daniel P. Siewiorek Associate Professor of Human-Computer Interaction at Carnegie Mellon University. As part of my visit, I heard from a few of her doctoral students. One of the presentations was by Tzu-Sheng Kuo, who analyzed a housing assignment algorithm deployed in a mid-sized US city. As Tzu-Sheng told me, AI is increasingly playing a role in public services. For example, the allocation of housing services is driven by an algorithm. This algorithm uses personal information to predict the likelihood of an individual making emergency room visits, being a mental health inpatient, or finding themselves in jail. Based on the likelihood of one or more of these distinct events, the algorithm provides a score, which a case worker then uses to determine the individual’s relative rank on a housing waitlist. One could argue that the tool is objective and not biased towards a certain race. It is also explainable – we know what factors lead to a higher or lower ranking.
This mid-sized city is not the only one which uses a similar algorithm. Articles such as “Who’s homeless enough for housing? In San Francisco, an algorithm decides” or “Will algorithmic tools help or harm the homeless?” address the effectiveness of these algorithms in San Francisco and Los Angeles. The common subtext is that the algorithm itself poses risks. Among them is labeling: to identify someone as high risk may ensure that they remain high risk. If an individual is deemed to be high risk, then they may not receive certain housing services, which may reinforce behaviors that make them appear even more high risk in the future. Also, in some social workers’ view, the score now replaces what they would do – talk to unhoused individuals to determine their unique situation and needs. Another concern raised is that such algorithms shrink the problem to suit the scarcity of resources as opposed to working to solve the problem comprehensively. In response to these concerns, in the first article mentioned above, Joe Wilson says, “No computer is going to help me decide the worth of another human being, and who gets what, when and how much.” Joe is the executive director of Hospitality House, a community-based organization in the Tenderloin in San Francisco that provides services for people experiencing homelessness.
In this case, the housing algorithm is highly explainable, but many professionals would call it unhelpful or even counterproductive. Even if the algorithm was perfect and the data was current and representative, all its other issues would remain. Can it really be considered a trustworthy AI system? Ultimately, trustworthiness is complex and situation specific.
Tax Notes chief correspondent Stephanie Soong discusses how countries are approaching implementation of the OECD’s two-pillar corporate tax reform plan and the future of the OECD after Pascal Saint-Amans’s departure. This …
Tax Notes chief correspondent Stephanie Soong discusses how countries are approaching implementation of the OECD’s two-pillar corporate tax reform plan and the future of the OECD after Pascal Saint-Amans’s departure.
This transcript has been edited for length and clarity.
David D. Stewart: Welcome to the podcast. I’m David Stewart, editor in chief of Tax Notes Today International. This week: pillars’ progress.
We’re checking in again on the OECD’s two-pillar project to reform the taxation of multinational companies. Despite ongoing challenges, the inclusive framework on base erosion and profit shifting (BEPS) continues to make progress toward a final design for implementation.
I’m joined by Tax Notes chief correspondent Stephanie Soong to talk more about where things stand now and where we expect them to go. Stephanie, welcome back to the podcast.
Stephanie Soong: Hey, thanks for having me.
David D. Stewart: Before we really dive into things, could you give our audience a bit of a refresher on what the two pillars are?
Stephanie Soong: Yes. The two pillars refers to a plan that the OECD brokered among 137 countries. It was meant to follow up on action 1 of the OECD base erosion and profit shifting project, which dealt with addressing the tax challenges of digital economy. The two pillars was basically a overhaul of the international tax system. It does two separate things.
Pillar 1 would revise profit allocation and nexus rules to give market jurisdictions a new taxing rate, called amount A, over some portion of multinational corporate residual profits that are tied to sales in a market jurisdiction. That’s where the consumers are located.
Pillar 1 also has amount B. That provides a fixed return for baseline marketing and distribution activities in market jurisdictions in line with the arm’s-length standard. It also includes dispute resolution and prevention mechanisms for tax certainty for amount A and issues related to amount A.
Pillar 1 is widely seen as part of the plan that’ll address the taxation of digital activity. It also requires countries to withdraw unilateral measures, like digital services taxes, that they may have adopted in the absence of a multilateral solution to tax digital activity and also prevents countries from introducing new such measures.
At the heart of pillar 2 are the the global anti-base erosion (GLOBE) rules, which are the income inclusion rule or the IIR and the UTPR, which is also known as the undertaxed payments or profits rule. This is a system of top-up taxation to make sure that companies do pay a 15 percent no tax rate.
Pillar 2 also includes the subject-to-tax rule. That’s a treaty-based rule under which source jurisdictions can impose a top-up withholding tax on some related-party payments that are taxed below a nominal rate of 9 percent.
As I’ve talked about before on this podcast, nearly all of the 140 some members of the inclusive framework had agreed in 2021 that they would adopt these two pillars.
David D. Stewart: This project has been going on for quite some time now. I understand there’s been some major changes since we last talked about it in July. Could you tell us what’s been going on?
Stephanie Soong: Yes. Not so much major changes of the project itself, but for the team behind it.
Everyone knows Pascal Saint-Amans was a director of the OECD Center for Tax Policy Administration (CTPA). He unexpectedly announced in September that he would step down as the director of the Centre for Tax Policy and Administration. At the end of October, he decided he would take a job elsewhere. Now, he’s at the Brunswick Group, which is a global advisory firm.
Pascal Saint-Amans, director of the OECD centre for tax policy and administration talks during a … [+] press briefing on June 7, 2017 prior to the signing by more than 60 countries of the Multilateral Convention to Implement Tax Treaty Related Measures to Prevent Base Erosion and Profit Shifting (BEPS) at the OECD headquarters in Paris during the annual OECD forum and OECD ministerial meeting. / AFP PHOTO / ERIC PIERMONT (Photo credit should read ERIC PIERMONT/AFP via Getty Images)
AFP via Getty Images
Grace Perez-Navarro, who was formerly deputy director, is now the director of the CTPA. That’s kind of interesting.
Pascal had been in this role for over 10 years and was the one who really got the BEPS project and the BEPS 2.0 project going. I asked him, “Why now?” Because it seems like a lot of people were asking him “Why now? It sounds like you’re going to jump, because it’s sinking.”
But he insisted that it was because he had been thinking about leaving for a while, but hadn’t had a chance. Now that the pillars seemed to be pretty stable and the work is still going, he felt it was time to go and get fresh blood in.
So Grace Perez-Navarro will be director until March next year. Then she’ll retire, and it’s still unclear who will take over as director after she leaves.
David D. Stewart: All right. For getting into the details, why don’t we start with pillar 2? As I mentioned, we talked about this back in July. What have been the updates since then?
Stephanie Soong: At the OECD, the secretariat and Working Party 11 are continuing to work on what they call a GLOBE implementation framework. They hope to get that out by the end of the year. This framework includes things like safe harbors, administrative guidance, standardized return forms, things that multinational enterprises, in scope, would need to comply with these rules once countries start adopting them. The OECD is also trying to get out the subject-to-tax rule — I think by the end of the year as well, if I’m not mistaken.
Meanwhile, countries are starting to move toward implementation, but it feels like “Who’s going to jump in the pool first?” No one really has adopted a global minimum tax at all except for the United States. Other countries are in varying stages of implementation. Some have started consulting on implementing pillar 2 in their respective countries. Others have produced draft legislation.
The EU comes to mind as the first to have a draft directive on the table. On November 17 the United Kingdom had their autumn statement. They had been consulting on how to implement pillar 2, and the chancellor of the exchequer, Jeremy Hunt, announced that the United Kingdom was indeed going to implement pillar 2 rules starting with the GLOBE.
He announced during the autumn statement that the United Kingdom would legislate for the IIR in the spring finance bill 2023. It also would adopt a qualified domestic minimum tax, which countries are allowed to adopt. These minimum taxes would take effect before the income inclusion rule. It’s an optional measure that countries can adopt under the GLOBE model rules.
The United Kingdom announced that it would also produce UTPR legislation. The rule takes effect no later than the accounted period, beginning on or after December 31, 2024, which really means the UTPR would take effect in 2025.
There are countries moving. It’s just a matter of who’s going to go first.
David D. Stewart: Before this project even got started, the United States had implemented a sort of minimum tax in the global intangible low-taxed income rules. Since we last talked, the United States has adopted a second sort of minimum tax, the book minimum tax, in the corporate alternative minimum tax as part of the Inflation Reduction Act. How does that factor into all of this discussion on pillar 2?
WASHINGTON, DC – AUGUST 16: U.S. President Joe Biden (C) signs The Inflation Reduction Act with … [+] (L-R) Sen. Joe Manchin (D-WV), Senate Majority Leader Charles Schumer (D-NY), House Majority Whip James Clyburn (D-SC), Rep. Frank Pallone (D-NJ) and Rep. Kathy Catsor (D-FL) in the State Dining Room of the White House August 16, 2022 in Washington, DC. The $737 billion bill focuses on climate change, lower health care costs and creating clean energy jobs by enacting a 15% corporate minimum tax, a 1-percent fee on stock buybacks and enhancing IRS enforcement. (Photo by Drew Angerer/Getty Images)
Getty Images
Stephanie Soong: I think a lot of the hope for the United States implementing pillar 2 rules really kind of rode on the reforms on GILTI. Because the IIR was sort of inspired by GILTI, and countries were hoping that the United States would be able to amend the GILTI so it’s more in line with the IIR.
But because now we’ve got this corporate AMT in place, the question is how will it work with the IIR under the GLOBE framework? How are other countries going to treat it? Is it going to be what they call a qualified IIR in line with the rules? Is it going to be another controlled foreign corporation regime?
Kim Clausing recently said at a conference it could be a bridge toward eventual GLOBE IIR adoption for the United States, but it’s still really unclear what’s going to happen now.
David D. Stewart: Going beyond the United Kingdom and United States potential implementation of this, what are we seeing in other countries? How are they approaching pillar 2?
Stephanie Soong: A lot of countries, like Switzerland, Canada, Singapore, and Jersey, are all consulting on how their respective countries should adopt the GLOBE rules.
The only block of countries that has actually produced any kind of draft legislation that is complete is the EU. The EU produced a draft directive in December 2021 on the GLOBE rules. But they’ve had a lot of trouble getting unanimity on this directive, because the EU needs all member states to be on board with a tax directive. Hungary is still a holdout.
That has been a bit of a problem for the commission and for the EU Council, because they really want to get pillar 2 and GLOBE off the ground. There are five EU countries, including Germany and France, who said that they would go ahead anyway if Hungary continued to hold out and do an enhanced cooperation type of thing. Enhanced cooperation means that a group of member states can go ahead with a tax directive as long as a minimum of nine are on board. We’ll have to wait and see what happens with that.
Meanwhile, the United States actually announced recently that it would cancel its treaty with Hungary, citing Hungary’s opposition to the GLOBE directive, the pillar 2 directive in the EU, and that has riled up some Republican lawmakers in the United States. A bunch of Republican lawmakers actually sent a letter to Treasury saying that the United States should reverse its stance on canceling the treaty with Hungary.
Never mind the fact that the Hungary-U.S. treaty is going to be canceled. It’s going to be canceled anyway, because there is a pending new Hungary-U.S. treaty that has not been approved yet in the Senate.
A lot of threads to unspool here, because pillar 2 is really the one pillar that has been ramping up in terms of action.
David D. Stewart: Well, speaking of action, what are we expecting next on pillar 2?
Stephanie Soong: I’m waiting for the OECD to release their implementation framework for the GLOBE rules as well as the STTR (subject-to-tax-rule). I am just waiting to see which other countries are going to be the first to pick a punch, really. We’ll see what happens.
vector map with location of the Organisation for Economic Co operation and Development for your … [+] project
getty
David D. Stewart: All right. Well, then — turning to pillar 1, what’s been happening since we last talked about that?
Stephanie Soong: Pillar 1 isn’t getting as much attention these days as pillar 2, just because pillar 1 seems to be more of a far-fetched prospect in terms of implementation, which is fair because a lot of the rules have not been finalized yet, even though countries have agreed to implement it. So the OECD has been in hard at work producing public consultation documents for draft amount A rules.
In July, they produced a public consultation draft on amount A rules and held a public consultation meeting in September in Paris. Then they produced another public consultation document on draft administration rules and tax certainty aspects of amount A and issues related to amount A. That consultation just closed, and comment letters have been published. I don’t know when that consultation meeting will happen, but I would expect that to take place at some point.
Meanwhile, G-20 finance ministers and leaders continue to push ahead with pillar 1 and pillar 2. Recently, the G-20 leaders had a meeting in Bali. In their declaration, they reaffirmed their commitment to both pillars and actually called on countries to sign on to the multilateral convention that’s needed to implement the pillar 1 amount A rules when those rules are ready.
The OECD is currently working on producing that multilateral convention. They’re hoping to get it ready for countries to sign by mid-2023. Meanwhile, we’re also waiting for other, what they call a building block from amount A, including a document on unilateral measures. This document is expected to discuss what kind of unilateral measures countries will be expected to withdraw once those amount A rules are in place.
David D. Stewart: Now, you alluded to this before, but pillar 1 does seem to be a much heavier lift than pillar 2. What can we expect if it fails?
Stephanie Soong: Well, we can expect lots of DSTs, probably. DSTs are like the boogeyman in the room. All countries don’t want DSTs, especially the U.S. Republicans and Democrats can agree on one thing and that’s that digital services taxes are bad. If pillar 1 fails, if amount A fails, then I’d probably bet that those DSTs will be back in full force, because a lot of countries are fed up with seeing what they see as digital giants and companies with digital services really just escaping tax. So they’re going to probably act on unilateral measures.
That brings the question up about whether we’ll see more trade disputes over digital services taxes. And already, we’re seeing some action toward that direction, because the USTR, the Office of the U.S. Trade Representative, has been consulting on the Kenya trade deal and with the head of public consultation.
Some trade groups have been bringing up this idea again that as a condition of the new Kenya-U.S. trade agreement, they should drop their DST. So now, we’re seeing trade and digital taxation creep back together and convene. I would expect more trade tensions to restart if amount A fails.
Also, the U.K. Public Accounts Committee is supposed to hold a hearing on the U.K. DST on December 8, just questioning HM Revenue and Customs and Treasury about what the plan is for withdrawing these taxes if pillar 1 amount A comes into effect.
So, you know, DSTs for everybody.
David D. Stewart: Well, then, stepping back for the bigger picture — is there any sense of when this entire project is going to be finished?
Stephanie Soong: This is a very good question. I would guess that this will never end, which is good for me, because I can write about it all day long. Seriously, I don’t know if there is an end in the project.
I guess, in the near future, the dates I can see maybe for the major parts of the project to be finished is 2024, 2025 maybe. But I think that it’s going to be an ongoing process. I don’t know if there’s going to be a real end date in sight, unless the G-20 decides that they want to scrap everything. I don’t think it’s going to happen. But I think it’ll be with us for a long time to come.
David D. Stewart: Well, that’s great, because I love having you on the podcast. And so we’re going to get to keep doing that for the foreseeable future. Stephanie, thank you for being here.