How should companies use AI? A review of the research team approaches of the US-based research institutes, and the case for responsible AI: The race to the bottom

Holger Hoos, a research assistant at RWTH Aachen University in Germany said that he is looking at a situation where the top-notch research is done within the research labs of a small number of mostly US-based companies.

Companies had invested more heavily in responsible-AI research in the past, says Vallor, but that interest waned with the boom in generative AI, prompting a “race to the bottom” to capitalize on the market. “The knowledge about responsible AI is all there, the problem is that large AI companies don’t have incentives to apply it,” she says. “But we could change the incentives.”

Companies that develop and deploy AI responsibly could face a lighter tax burden, she suggests. Vallor asserts that those who do not want to adopt standards should be compensated for the harm they cause.

For that scrutiny to happen, however, it is imperative that academics have open access to the technology and code that underpins commercial AI models. “Nobody, not even the best experts, can just look at a complex neural network and figure out exactly how it works,” says Hoos. “We know very little about the capabilities and limitations of these systems, so it is absolutely essential that we know as much as possible about how they are created.”

Theis says that many companies want to be able to work with more people because they want open access for their models. “It’s a core interest for industry to have people trained on their tools,” he says. Meta, the parent company of Facebook, for example, has been pushing for more open models because it wants to better compete with the likes of OpenAI and Google. The flow of new ideas will be made easier by the fact that people are able to look at the models.

But it is unrealistic to expect that companies will give away all of their “secret sauce”, says Hoos — another reason it is important that academia retains the capability, in both technology and talent, to keep up with industry developments.

The different approaches of industry and academic researchers have been studied by the group. They analysed papers presented at a variety of AI conferences between 1995 and 2020 to see how the composition of a research team was related to the novelty of their work, and its impact in terms of citations and models created.

Artificial Intelligence Research in Canada: How to Make the Most of the Public Funding and Support Its Universities? The Case of the CLAIRE Lab

To make the most of that freedom, however, academics will need support — most importantly in the form of funding. Theis says that a strong investment into basic research would be useful.

Although governments are unlikely to be able to match the huge amounts of money being splashed around by industry, smaller, more focused investments can have outsized influence. Canada has an effective artificial intelligence strategy that has not cost a lot of money. The country has invested around Can$2 billion (US$1.46 billion) in AI initiatives since 2016, and in 2024 announced plans to spend another Can$2.4 billion over the next few years. Much of the money is earmarked for providing university researchers with access to the computing power they need for applications that useArtificial Intelligence, as well as to support Responsible AI research and to recruit and retain top talent. Canada is able to punch above it’s weight and remain near the top of the global rankings in both academic research and commercial development because of this strategy. It placed 7th in the world for Nature Index output in AI research in 2023, and 9th in natural sciences overall.

The confederation of laboratories for artificial intelligence research in Europe, known as the CLAIRE, has put forward a plan that is even more ambitious. The plan is inspired by the sharing of large, expensive facilities in the physical sciences. “Our friends the particle physicists have the right idea,” says Hoos. “They build big machines funded by public money.”

The data sets that companies have access to are larger than the ones used for training models, as users interact with them. When it comes to training large language models for natural language processing, it is going to be difficult for academia to keep up.

Environmental Science and Technology Policy: Working with Fishing Crews and the Community to Monitor Drought in Northern U.S. During the Cold War

So I started engaging with the fishing crews and their families directly. I attended their meetings to hear about the drought’s impacts. Listening to the community informed my research, and it led to me becoming a strong advocate for scientists who do work that can be applied directly to society. I make sure that they engage with the intended beneficiaries of their research early on, and on a continuous basis.

In 2009, after my postdoc, I took a two-year policy fellowship through the American Association for the Advancement of Science that placed me with the US National Oceanic and Atmospheric Administration’s (NOAA) Climate Program Office, Silver Spring, Maryland. There, I worked to build partnerships across NOAA on climate and coastal and marine ecosystems, including an initiative to establish a regional drought early-warning system for the southeastern United States. I was then hired as an ecosystem-science adviser at the office, working on cross-agency climate-adaptation and -resilience activities and programmes. I have taken assignments at the White House office of science and technology policy twice during my career. Starting in 2014, I worked at the OSTP for three years, helping officials to implement then-president Barack Obama’s Climate Action Plan. I return to the OSTP in 2021. I am now the chief of staff for the Climate and Environment Team and the assistant director for climate resilience. I got hooked in the science-policy world and I never left.

We are working with government agencies to develop nature-based solutions. I am trained as an ecologist and understand the science behind that. For example, restoring a marsh would also buffer nearby communities, buildings or roads against sea-level rise or storm surges. Such initiatives both strengthen nature and provide protective benefits for people.

I became interested in policy when I was pursuing my master’s degree in astronomy and astrophysics at the University of Victoria, Canada. A good friend encouraged me to be vice-president of the graduate student association, through which I eventually helped to negotiate dental coverage for members. It was during my meeting with the university provosts that I realized what needed to be done. So many problems needed solving through policy change. I thought, ‘I actually really enjoy this.’

I now work at the National Sciences and Engineering Research Council (NSERC), which distributes government funds to university researchers throughout Canada. My job is to collect data to see whether council policies are working.

I work for the chief data officer and am responsible for all data related infrastructures for the NCSL, which includes data stewardship, analysis and coordination with other government agencies and ministries.

The Canadian government is investing money into electric vehicles, their batteries and recycling them. If someone in parliament says, “Electric vehicles are the future. We have to look through reports to find projects funded by the NSERC. It takes a lot of time. We are attempting to incorporate machine learning into our processes so that you don’t need to read every report with the phrase engineering as a title. With machine learning, you can get your hands on a larger data set. We can help officials craft policy more efficiently by using those data.

Source: Science-policy advisers shape programmes that solve real-world problems

Accessing data on the people who apply for grant money to understand equality, diversity and inclusion in science-policy advisers shape programmes that solve real-world problems

One of the areas we’re working on is equality, diversity and inclusion (EDI). I have attention deficit hyperactivity disorder, am on the autism spectrum and have mental-health challenges, so I want the policies to be more effective. I had trouble in my science education because of these things. But I know I’m lucky to have made it to where I am.

Our office put forward the idea that we needed to collect data on the people who get grant money to better understand EDI. There was data collected here and there. I and anNSERC colleague used that data to analyse it.

We showed places that the council is doing well but where there are gaps, as well as aspects in which the council is doing poorly. For example, we found that the diversity of people applying for funding is not yet representative of the Canadian population as a whole. Census data shows how many applications are needed to represent all award recipients. We would expect 50 black people to apply for a fellowship if the pool of students were representative of the population, but we received only 30 applications.

For example, the NSERC can say to a university, “We see that you didn’t report having any Black applicants to your PhD programmes last year. Are you okay with that? Or, “Why might that be?” We know about these gaps, because we can see the data.

Source: Science-policy advisers shape programmes that solve real-world problems

Models of carbon dioxide emission in the Indian economy: How to reach net-zero by 2060, 2070, 2080 and 2100?

For instance, India’s Ministry of Environment, Forest and Climate Change reached out to academics and think tanks involved in modelling. Our group came up with scenarios discussing what it would take for the country to reach net-zero carbon dioxide emissions by 2060, 2070, 2080 and 2100. The models gave ministers potential paths that the country could take to reach net zero, and these informed India’s decision to commit to doing so in the Glasgow Climate Pact at COP26, the 2021 UN climate change conference.

It’s very difficult to get an audience with policymakers. It’s necessary for us to follow up several times. I don’t say anything. They lap it up. That never happens. It takes many conversations.

To discuss what the biggest impact is, you have to speak their language. If you’re talking to a legislative member who works directly with people, then it could be social impacts. By contrast, for policymakers at the highest levels of government, it could be economic impacts and investments. These are the big tickets that resonate well. At the end of the day, they want to make a difference.

Source: Science-policy advisers shape programmes that solve real-world problems

Science & Policy Advisors Shape Programmes That Solve Real-World Problems: 25 Years in Bengaluru as a Forest Ecologist

I spent 25 years in the Indian Institute of Science in Bengaluru as a forest ecologist. I was monitoring a biodiversity hot spot in India.

Originally, I assumed I would be a tenure-track faculty member at a research university. I was doing research as a postdoc at the Florida State University that I thought would benefit the community. I was looking at the causes of declines in oyster populations.

In policymaking, deep knowledge of a subject doesn’t always make you an effective science adviser. Having the ability to build trust with policymakers is more important than just having aholistic understanding of an issue.

Alongside running my research group — an international team working on nanotechnology for sustainable chemistry and clean energy solutions — I volunteer at organizations such as the World Economic Forum (WEF) and the International Union of Pure and Applied Chemistry (IUPAC), helping global leaders to create more effective public policies that are backed by sound evidence.

Our reports have identified emerging technologies that will shape the future of science. Some of the predictions have come true. The 2015 report identified the gene-editing tool CRISPR–Cas9 as a revolutionary technology. The scientists who discovered CRISPR won a Nobel prize five years later.

The report highlighted thevaccines that are made from mRNA. We didn’t know about the COVID-19 pandemic at the time. We think that people are not paying enough attention to this technology, and that the governments should put more resources into developing it.

It has been found that scientific policy can help to steer research in this area. An innovative chemistry is needed that can be read by machines. We are creating a standard way to communicate information about chemical substances. This will accelerate the implementation of AI in scientific discovery.

Systems thinking also means connecting various disciplines. In the case of chemistry, it is always putting people at the center because it entails connecting the nature of compounds with their role in health, the economy, and the environment.

Source: Science-policy advisers shape programmes that solve real-world problems

The Future of Artificial Intelligence: From the States to the Global, Beyond the Restriction and Regulated Regime: The Case of the Connecticut AI Law

We’re working with governments worldwide to provide guidelines, teaching tools and training workshops. In the past five years, we’ve run workshops in South Africa, the United States and Egypt, training secondary-school teachers to bring this approach to their classrooms.

This requires producing high-quality reports that stand up to the most rigorous scrutiny. But beyond the report, you need to build trust through active listening and empathy. It begins on the day you are asked to advise, and goes until the day the legislation is implemented.

Scientific advice is mainly about providing the best current knowledge in context. Just as decision makers work with lawyers to ensure their decisions are constitutionally sound, they need to work closely with scientific advisers to incorporate the latest knowledge into policies.

As with the GDPR, there might be some cases where compliance with EU rules makes business sense for US firms, but it would mean the United States would be left overall less regulated, meaning that individuals will be less protected from AI abuses. The core of the Act was intact despite being the subject of a fair amount of lobbying. We have to wait and see if the US state laws stay the course.

For these reasons, lobbying groups claim to prefer national, unified AI regulation over state-by-state fragmentation, a line that has been parroted by big tech companies in public. But in private, some advocate for light-touch, voluntary rules all round, showing their dislike of both state and national AI legislation. If neither kind of regulation emerges, AI companies will have preserved the status quo: a bet that two divergent regulatory environments in the EU and United States — with a light-touch regime in the latter — favour them more than the benefits of a harmonized, yet heavily regulated, system.

The tech industry’s power can extend beyond this type of passive inspiration. The section on generative artificial intelligence was removed from the draft bill in Connecticut after being pushed by industry. The bill received support from a few tech companies, but it still is in limbo. The governor of Connecticut threatened to veto the bill, due to the associations’ claims that it would stifle innovation. Many of the more comprehensive Artificial Intelligence bills are being considered by a number of states. The bill is expected to be altered to prevent it from having a negative impact on innovation.

A major difference between the state bills and the AI Act is their scope. The risk-based system set up by the act is meant to protect fundamental rights, but some uses of the tech, such as social scoring, are not allowed. High-risk AI applications, such as those used in law enforcement, are subject to the most stringent requirements, and lower-risk systems have fewer or no obligations.