Search the archive.

Despite efforts by several countries to establish biobanks that collect phenotypic and genetic informations, the creation of biodiversity biobanks that collect nonhuman samples is still lagging behind.1,11,12 In a 2016 publication, a PubMed search for the keywords “biobank,” “biobanking,” “genebank,” and “gene bank” revealed that only 9 of 498 search results referred to animal biobanks.13 The International Society for Biological and Environmental Repositories' (ISBER) official journal, Biopreservation and Biobanking, even released a call for more manuscript submissions related to environmental biospecimen biobanking.14

Biodiversity biobanks should be established and supported globally as a response to the expected loss of global biodiversity, particularly in areas with strong human–wildlife interactions and habitat modification.15–17 However, one study noted that even biologically megadiverse developing countries (MDCs) lack infrastructures to maintain biological collections.12

With each species that disappears, vast amounts of information about their biology, ecology and evolutionary history is irreplaceably lost. The Frozen Ark Project (www.frozenark.org), a charity with headquarters at the University of Nottingham, is one of the many organisations that are working hard to halt the loss of biodiversity by ensuring that all biological material that has been collected from threatened animal species is properly curated and used responsibly and sustainably in conservation, management and research. We aim to achieve this goal by: i) coordinating global efforts in animal biobanking; (ii) sharing expertise; (iii) offering help to organisations and governments that wish to set up biobanks in their own countries; (iv) providing the physical and informatics infrastructure that will allow conservationists and researchers to search for, locate and use this material wherever possible without having to resample from wild populations.

The origins of the Ark

To maximise the potential of biobanking, standardised sampling procedures and preservation techniques need to be developed, and stored samples should be supported by an accessible database. Many museums store animal material and university laboratories and zoos have collections of tissue, DNA, banks of cells and gametes, mostly used for research purposes. However, many tissue collections are not in a form suitable for the long-term preservation of undamaged DNA and the role of cryopreserved biomaterial in UK conservation programmes has been growing. The Frozen Ark and its partner biobanks receive a high volume of requests for samples, but none possess the infrastructure to actively quantify this demand, emphasising the current lack of organisational infrastructure in many collections.

There has been a marked rise in cooperative international efforts to protect and preserve the global environment in the last few decades.[1] This trend is evidenced by agreements such as the 1972 Stockholm Declaration on the Human Environment, the 1987 Montreal Protocol banning the use of Chlorofluorocarbons (CFCs), the 1997 Kyoto Protocol seeking to reduce greenhouse emissions and combat climate change, and its successor, the Paris Agreement. Moreover, the need for international cooperation in environmental protection and regulation becomes even more salient when one considers the existential threat of anthropogenic climate change and the potentially catastrophic consequences that humanity will face in the twenty-first century and beyond. According to a United Nations (UN) report, despite continued effort by international bodies at setting global standards for environmental protections and greenhouse gas reduction, these international environmental laws are facing a problem[2] that is relatively common to international legal frameworks—lack of sufficient enforcement methods.[3]

In January of 2019, the UN released a global assessment on the environmental rule of law, the first ever report of this kind.[4] What they found was that despite a substantial increase in the amount of environmental protection agencies and laws, widespread failure to adequately enforce regulations has impeded the international effort to combat numerous environmental threats.[5] While the news issued by this report is undoubtedly concerning for the future of the planet, it is unlikely to be surprising to those familiar with the history and operation of international law. Lack of sufficient enforcement mechanisms is an issue that has affected international bodies and agreements throughout the twentieth century and this issue continues to frustrate the international community’s efforts to implement effective global policy in the twenty-first century.[6]

MONTREAL, Dec 9 (Reuters) - This month's  aim to hammer out a new, global agreement. But this isn't the first time governments have tried to halt environmental loss through agreed targets.

In fact, many of the 24 conservation targets under discussion now aim to avoid past mistakes and improve on the world's last set of conservation goals - the .

The science on the status and trends of biodiversity around the world has highlighted the crisis that nature and humanity face. Today, approximately one million species are at risk of extinction globally, the majority of land and seas have been significantly modified by human activity, populations of wild species continue to decline, and ecosystem services—from disease buffering to pollination—are at risk of loss (IPBES, ; WWF, ). The result is reflected in the recent review of progress toward achieving the 2020 Aichi Targets of the U.N. Convention on Biological Diversity—none of the targets were achieved and some have regressed (Secretariat of the Convention on Biological Diversity, ). Through the Endangered Species Act (ESA), the U.S. employs one of the world's most powerful laws for biodiversity conservation. The ESA aspires to support biodiversity by preventing extinction, recovering imperiled species, and protecting the ecosystems in which they live. Now 50 years old, the goals of the ESA have yet to be fulfilled: hundreds of species listed under the ESA continue to decline (Evans et al., ; Gratwicke et al., ; Malcom et al., ; Schwartz, ), and many hundreds of species not yet listed continue to decline to the point that ESA protections may be required. Despite being credited with preventing numerous extinctions (Gerber et al., ), the ESA is but a single law among many that expresses Congressional intent for protecting nature. While biodiversity protection benefits from high public support (Bruskotter et al., ; Burivalova et al., ; Pollack, ), a complex, and often contradictory policy landscape, the lack of coordination among agencies, and polarization of public perception about biodiversity conservation (Malcom & Li, ) all contribute to a general failure of the U.S. to enunciate a clear vision for the future of its biodiversity and demonstrate progress toward this goal. To end this, a national Strategy that reflects the plurality of values for biodiversity conservation and societal expectations of government programs on behalf of nature is warranted.

National-level decisions for addressing the biodiversity crisis also have global implications, particularly since the U.S. has important historical standing as a global leader in environmental protection. Whether it is direct intervention through providing resources and direction to curb wildlife trafficking (U.S. Congress, ) or participating in multilateral discussions and negotiations, U.S. actions carry weight and influence across the globe.

An enduring problem with today’s generative artificial intelligence (AI) tools, like ChatGPT, is that they often confidently assert false information. Computer scientists call this behavior “hallucination,” and it’s a key barrier to AI’s usefulness.

Hallucinations have led to some embarrassing public slip-ups. In February, AirCanada was forced by a tribunal to honor a discount that its customer-support chatbot had mistakenly offered to a passenger. In May, Google was forced to make changes to its new “AI overviews” search feature, after the bot told some users that it was safe to eat rocks. And last June, two lawyers were fined $5,000 by a U.S. judge after one of them admitted he had used ChatGPT to help write a court filing. He came clean because the chatbot had added fake citations to the submission, which pointed to cases that never existed.

Artificial Intelligence (AI) is unlikely to destroy humanity, but it is already creating societal upheavals across industry, government, education and the creative arts. 'Deepfakes' are already being blamed for phoney political media. AI has arrived and we need to regulate it. But, regulating AI is difficult because, like many technologies, AI is neither inherently good nor bad. It depends on how it is used.

The EU recently issued rules on AI, segregating its use into tiers. These span unacceptably high-risk and banned uses down to minimal-risk and lightly regulated applications. While this is a useful framework and easy to conceptualize, determining risk levels isn’t easy. The best regulation focuses not on the AI tools, but on their users and usage. We apply this same method to the regulation of knives. Up-front regulations aren’t going to be enough. Instead, we need to continuously examine AI’s outcomes.

There are several arguments against excessive AI regulation – and for some, against any regulation at all. 

Innovation and Progress

In a comparable manner, there is a first mover advantage to regulation. Thanks in large part to the interconnected nature of 21st century networks, the government that establishes the first set of rules defines the discussion from that point forward for all nations. The classic example of this is the European Union’s 2108 General Data Protection Regulation (GDPR), which has become the standard for privacy policy around the world.

Once again it appears as though the EU, which has been in the lead in establishing digital platform policy with its Digital Markets Act and Digital Services Act, is also in the lead on establishing AI policy. On June 14 the European Parliament overwhelmingly approved the AI Act. Following its adoption, the regulatory machinery of the European Commission will begin developing enforceable policies.

To keep the corporate AI race from becoming reckless requires the establishment and development of rules and the enforcement of legal guardrails. Dealing with the velocity of AI-driven change, however, can outstrip the federal government’s existing expertise and authority. The regulatory statutes and structures available to the government today were built on industrial era assumptions that have already been outpaced by the first decades of the digital platform era. Existing rules are insufficiently agile to deal with the velocity of AI development.

Former Google Executive Chairman and current AI evangelist Eric Schmidt has warned, “There’s no one in government who can get it [AI oversight] right.” While Mr. Schmidt recognizes the need for behavioral expectations, his solution is, “I would much rather have the current companies define reasonable boundaries.” Such a self-regulatory approach is the same kind of “leave us alone” solution that has been championed by digital platform companies for the last 20 years. The results of this strategy speak for themselves in well-known current online harms, such as the unprecedented invasion of personal privacy, market concentration, user manipulation, and the dissemination of hate, lies, and misinformation. AI demands something better than corporate self-regulation when we know the chase for profits is likely to outstrip the implementation of meaningful guardrails.

It’s official. After three years, the AI Act, the EU’s new sweeping AI law, jumped through its final bureaucratic hoop last week when the European Parliament voted to approve it. (You can catch up on the five main things you need to know about the AI Act with this story I wrote last year.)

This also feels like the end of an era for me personally: I was the first reporter to get the scoop on an early draft of the AI Act in 2021, and have followed the ensuing lobbying circus closely ever since.

As lawmakers and regulators in the U.S. consider policy born of their Big Tech concerns such as data privacy and artificial intelligence, they should carefully consider how such changes could end up trampling the small and midsize businesses that drive innovation and competition.

While policymakers may have Google and Facebook in mind, the actual policies could unintentionally create new regulatory burdens that could deter investment in smaller businesses and prevent new companies from emerging. For example, calls to end Section 230 — part of a 1996 law that protects internet companies from some lawsuits — portray it as a handout to Big Tech, when in practice it would mean new social media companies would face liability early on, making it more difficult to compete and discouraging them from carrying user-generated content that provides new opportunities or ways of connecting.

New York City isn’t alone in this struggle. As policymakers around the world have attempted to create guidance and regulation for AI’s use in settings ranging from school admissions and home loan approvals to military weapon targeting systems, they all face the same problem: AI is really challenging to define.

Subtle differences in definition—as well as the overlapping and loaded terminology different actors use to describe similar techniques—can have major impacts on some of the most important problems facing policymakers. Researchers typically refer to techniques that infer patterns from large sets of data as “machine learning,” yet the same concept is often labeled “AI” in policy—conjuring the specter of systems with superhuman capabilities rather than narrow and fallible algorithms. And some technologies commercially marketed as AI are so straightforward that their own engineers would describe them as “classic statistical methods.”

The need to distinguish real from synthetic media and factual content from ‘hallucinations’ is essential for maintaining the shared factual foundations underpinning social cohesion. Advances in generative AI have made it more difficult to distinguish between AI-generated media and real images, audio, and video recordings. Already we have seen AI-generated voice technology used in financial scams.29

Creators of the most powerful AI systems have acknowledged that these systems can produce convincing textual responses that rely on completely fabricated or out-of-context information.30 For society to absorb these new technologies, we will need effective tools that allow the public to evaluate the authenticity and veracity of the content they consume.

The private sector under-invests in research that ensures that AI systems are safe and secure. Despite nearly USD 100 billion of private investment in AI in 2022 alone, it is estimated that only about 100 full-time researchers worldwide are specifically working to ensure AI is safe and properly aligned with human values and intentions.24

In recent months, companies developing the most powerful AI systems have either downsized or entirely abolished their respective ‘responsible AI’ teams.25 While this partly reflects a broader trend of mass layoffs across the technology sector, it nonetheless reveals the relative de-prioritization of safety and ethics considerations in the race to put new systems on the market.

Commercial actors may not have sufficient incentives to protect their models, and their cyberdefense measures can often be insufficient. In early March 2023, Meta demonstrated that this is not a theoretical concern, when their model known as LLaMa was leaked to the internet.22 As of the date of this publication, Meta has been unable to determine who leaked the model. This lab leak allowed anyone to copy the model and represented the first time that a major tech firm’s restricted-access large language model was released to the public.

Watermarking of AI models provides effective protection against stealing, illegitimate redistribution and unauthorized application, because this practice enables legal action against identifiable leakers. Many digital media are already protected by watermarking – for example through the embedding of company logos in images or videos. A similar process23 can be applied to advanced AI models, either by inserting information directly into the model parameters or by training it on specific trigger data.

AI systems present a unique challenge in assigning liability. In contrast to typical commercial products or traditional software, AI systems can perform in ways that are not well understood by their developers, can learn and adapt after they are sold and are likely to be applied in unforeseen contexts. The ability for AI systems to interact with and learn from other AI systems is expected to expedite the emergence of unanticipated behaviors and capabilities, especially as the AI ecosystem becomes more expansive and interconnected.

Several plug-ins have already been developed that allow AI systems like ChatGPT to perform tasks through other online services (e.g. ordering food delivery, booking travel, making reservations), broadening the range of potential real-world harms that can result from their use and further complicating the assignment of liability.19 OpenAI’s GPT-4  system card references an instance of the system explicitly deceiving a human into bypassing a CAPTCHA bot-detection system using TaskRabbit, a service for soliciting freelance labor.20

AI is developing at a breakneck pace and governments need to catch up. The establishment of AI regulatory agencies helps to consolidate expertise and reduces the risk of a patchwork approach.

The UK has already established an Office for Artificial Intelligence and the EU is currently legislating for an AI Board. Similarly, in the US, Representative Ted Lieu has announced legislation to create a non-partisan AI Commission with the aim of establishing a regulatory agency. These efforts need to be sped up and taken up around the world.

At present, the most advanced AI systems are developed through training that requires an enormous amount of computational power – ‘compute’ for short. The amount of compute used to train a general-purpose system largely correlates with its capabilities, as well as the magnitude of its risks.

Today’s most advanced models, like OpenAI’s GPT-4 or Google’s PaLM, can only be trained with thousands of specialized chips running over a period of months. While chip innovation and better algorithms will reduce the resources required in the future, training the most powerful AI systems will likely remain prohibitively expensive to all but the best-resourced players.