To realize A.I.’s potential for health and science, we need to build smart guardrails.

By Ray Chambers and Michelle Williams ?

The recent advances in the field of A.I. have ignited vociferous debate around the risks vs. rewards of these transformative technologies. There is no doubt that A.I. will revolutionize the way we communicate, innovate, govern, and structure our societies. This seismic shift is, in fact, already underway. Confronted with the breathtaking power of these tools, experts, policymakers, and industry leaders alike are calling for urgent and thoughtful regulation to encourage advancement while preventing catastrophic downstream consequences.?

Earlier this month,?the White House convened?the CEOs of?four leading A.I. developers – Sam Altman of OpenAI, Dario Amodei of Anthropic, Satya Nadella of Microsoft, and Sundar Pichai of Alphabet – “to underscore that companies have a fundamental responsibility to make sure their products are safe and secure before they are deployed or made public.”

In his recent testimony to Congress, Altman acknowledged these safety concerns and enumerated risks ranging from "interactive disinformation" to cyberattacks to the widespread proliferation of autonomous weapons. In subsequent comments to the press, he outlined additional threats from the convergence of A.I. tools with synthetic biology, including?the creation of novel pathogens.

A.I. represents?a new frontier?in the field of synthetic biology – one with the potential?to “turbocharge” existing risks, such as increasing the ability of bad actors to engineer pathogens and bioweapons. A.I. could, for example, “upskill” people with adjacent expertise, or provide step by step instructions to troubleshoot lab challenges and achieve complicated end results.?

Experts?have warned that A.I. can?create more dangerous viruses by identifying pathogen mutations that enhance transmissibility or lethality. It could also be used to alter viral proteins, increasing their resistance to vaccines and other countermeasures. Similarly, by suggesting small changes to genetic structures, bad actors could use these technologies to evade detection by existing DNA synthesis screening tools.

These risks are compounded by the ability of A.I. to produce disinformation and misinformation. Indeed, a sophisticated effort to manipulate reality could make it nearly impossible for authorities to mount an effective response to an outbreak – whether manmade or naturally occurring.?

This week,?a small group of leaders from within industry, government, and academia are being convened by the problem-solving organization Helena at the The Rockefeller Foundation 's Bellagio Center. A primary focus of their meeting will be to explore critical interventions to increase global biosafety and biosecurity. Their discussions will consider the role of national governments, multilateral entities, the scientific community, and the private sector.??

As it pertains to A.I., the group’s agenda will include conversations on risks as outlined above, as well as the ways in which mis- and disinformation campaigns could sow chaos in the event of biosecurity incidents like?the recent military seizure of a public health lab in Sudan. In the current geopolitical climate, such occurrences could be leveraged to stoke?fear and uncertainty, cast blame on adversaries, or incite further violence.

If this group can arrive at specific recommendations, they will be well-positioned to work with leaders across Europe and Africa, in the UK and U.S., and within and between multinational entities to develop cohesive policies at the nexus of A.I. and synthetic biology.

The World Health Organization is also well-positioned to establish its authority on novel bio risks by integrating emerging threats from A.I. into their own approaches and by helping developing countries build guidelines and capacity. In tandem, WHO might collaborate with the UN Security Council and UN Secretary General’s Mechanism (UNSGM) to jointly evolve protocols and enable a nimble response in the event of a biological crisis – whether the crisis appeared in a war zone like Sudan or in an economic center.?

In his testimony to Congress, Altman warned that few policymakers understand novel technologies at the level required to thoughtfully regulate them. His comments highlighted the need to embed advanced technical expertise into our governing institutions. Leaders might consider creating??“A.I. Councils” to develop informed regulatory frameworks??and align across national borders on consistent and coordinated approaches less vulnerable to?regulatory arbitrage.?

However the conversations in Bellagio unfold, the time to act is now. The intersection of A.I. and scientific discovery will furnish incredible breakthroughs. It is already showing promise in facilitating?early cancer detection,?accelerating the development of vaccines and therapeutics?and designing compounds with the potential to become?new superbug-killing antibiotics. These are just a few examples in a rapidly expanding universe of opportunities.?Such innovations hold the potential to alter the course of public health and should be resourced and encouraged – but only with appropriate guardrails in place.?

Special thanks to Claire Qureshi and her colleagues at Helena.