Governments must act if AI is to deliver for humanity. While we wait, investors must step up and work with companies to define new guardrails.
Uncontrolled AI poses risks to humanity
On 12 September 2026, Dario Amodei, co-founder and chief executive of Anthropic, published an essay arguing that the pace of frontier model development must be set by the pace at which safety can be made reliable, rather than the other way around.
The essay, also supported by the chiefs of OpenAI, SpaceX and Google DeepMind, sets out three steps to ensure such pace is attainable. The first is voluntary and sits at company level: frontier developers should commit publicly to embedding external evaluators inside their organisations. This means giving access to workspaces, tools and permissions broadly comparable to those of internal risk assessment staff, and with a contractual right to publish findings on risk levels, incidents and practices without editorial control by the company. The second and third steps call for coordinated action by governments, first domestically and then internationally.
The underlying logic is powerful and urgent. Voluntary restraint by a small number of laboratories cannot hold indefinitely against competitive pressure. If AI is to serve humanity rather than endanger it, governments must regulate it.
Investors should play their part
Long-term investors, like society at large, have a clear interest in AI succeeding. Success means technology that enhances societal wellbeing. It does not mean technology that erodes it, whether through malicious misuse, loss of human control, or the concentration of decision-making power in systems that no one can inspect or contest.
As long-horizon investors in global markets, we cannot diversify away systemic harm. A catastrophic failure at one developer would not be contained to that holding. It has the potential to generate self-reinforcing cascading effects across economies, society and the environment, undermining long-term financial security and wellbeing for all. Safety is therefore not a constraint on the investment case for AI. It must sit at its core.
Government action may come too late
Coordinated government action will take time and will not be easy. The current US administration has set itself against binding federal regulation of AI and has discouraged state-level rules. Competition between the US and China is used to frame safety requirements as a self-imposed handicap. Where rules do exist, such as the European Union’s AI Act, implementation is contested and enforcement is untested. The likely outcome is regulation that is fragmented, slow and reactive, outpaced by the capability it was meant to govern.
Investors need a shared agenda
Investors are not bystanders in this. They own many of the companies developing and deploying these systems, and they appoint the boards that set strategy and hold executives to account. Where governments are slow, investors have the ability to push companies to establish guardrails across a sector that is otherwise struggling to act in society’s long-term interests.
For this to work, investors must act together to align behind a common set of expectations, providing developers with a coherent set of demands, rather than a patchwork. Critically, by establishing minimum investor expectations, we can help move the market. A company choosing to move first should not be disadvantaged for doing so.
Our asks
Building on Amodei’s three-point plan, we call on asset owners and managers to coalesce around the following five expectations for companies:
Pace the frontier. Commit publicly to tying scaling decisions to defined capability thresholds, with pre-agreed alignment, evaluation and safeguard requirements met before the next step is taken.
Build common standards. Work with peers, independent researchers, standards bodies and civil society on shared definitions of the capability thresholds that trigger enhanced review, common evaluation protocols, comparable reporting formats, and agreed competence and independence criteria for third-party evaluators [1]. Companies should disclose which standard-setting processes they take part in, what role they play in governing them, and where their own practice departs from the resulting standards.
Open the laboratory. Grant independent external evaluators continuing access sufficient to verify practices in training pipelines as well as in released models, on contractual terms that permit publication of key findings without company editorial control, and disclose those terms, including any redaction rights.
Report incidents. Identify safety and alignment incidents, including those arising from a company’s own use of AI systems in research and development, escalate them to the board, and report them externally.
Support regulation, publicly and consistently. Advocate for proportionate regulation of frontier systems, ensure that all lobbying is consistent with that position, including that of trade associations, committing to exit any that diverge.
To date, the role of investors in driving responsible AI has been neglected. While, key AI labs remain in private hands, the largest are public or on the route to a listing [2]. This means that these companies are, or will soon be, owned by the public they are there to serve. This vital shareholder voice needs to be heard. This ‘call to action’ seeks to begin to fill this gap. It is time asset owners and asset managers step up and fulfill their responsibility as responsible long-term stewards of their clients’ companies.
We invite fellow investors to join this call to action by contacting Julia Shatikova at julia.shatikova@sarasin.co.uk.
[1] This work should build on what exists, including the ISO/IEC 42001 management system certification, the NIST AI Risk Management Framework, the international network of AI safety institutes, and the evaluation practice being developed by organisations such as METR, mentioned in Amodei’s essay, whose Frontier Risk Reports assess developers and their models against capability and catastrophic risk criteria.
[2]Anthropic and Open AI are in the process of launching Initial Public Offerings of their shares.
Important information
This document is intended for retail investors and/or private clients. You should not act or rely on any information contained in this document without seeking advice from a professional adviser.
This is a marketing communication. Issued by Sarasin & Partners LLP, 50 George Street, London W1U 7DY. Registered in England and Wales, No. OC329859. Authorised and regulated by the Financial Conduct Authority (FRN: 475111).
Website: www.sarasinandpartners.com. Tel: +44 (0)20 7038 7000. Telephone calls may be recorded or monitored in accordance with applicable laws.
This document has been produced for marketing and informational purposes only. It is not a solicitation or an offer to buy or sell any security. The information on which the material is based has been obtained in good faith, from sources that we believe to be reliable, but we have not independently verified such information and we make no representation or warranty, express or implied, as to its accuracy. All expressions of opinion are subject to change without notice. This document should not be relied on for accounting, legal or tax advice, or investment recommendations. Reliance should not be placed on the views and information in this material when taking individual investment and/or strategic decisions.
Capital at risk. The value of investments and any income derived from them can fall as well as rise and investors may not get back the amount originally invested. If investing in foreign currencies, the return in the investor’s reference currency may increase or decrease as a result of currency fluctuations. Past performance is not a reliable indicator of future results and may not be repeated. Forecasts are not a reliable indicator of future performance.
Neither Sarasin & Partners LLP nor any other member of the J. Safra Sarasin Holding Ltd group accepts any liability or responsibility whatsoever for any consequential loss of any kind arising out of the use of this document or any part of its contents. The use of this document should not be regarded as a substitute for the exercise by the recipient of their own judgement. Sarasin & Partners LLP and/or any person connected with it may act upon or make use of the material referred to herein and/or any of the information upon which it is based, prior to publication of this document.
Where the data in this document comes partially from third-party sources the accuracy, completeness or correctness of the information contained in this publication is not guaranteed, and third-party data is provided without any warranties of any kind. Sarasin & Partners LLP shall have no liability in connection with third-party data.
© 2026 Sarasin & Partners LLP. All rights reserved. This document is subject to copyright and can only be reproduced or distributed with permission from Sarasin & Partners LLP. Any unauthorised use is strictly prohibited.
