
Investors call on chemical firms to phase out toxic substances
The group of 43 investors, with four trillion dollars in assets under management, warned that the sector is not acting fast enough to protect vital ecosystems and human health.
In a statement released on Thursday, they urged companies, including agrochemical producers, to commit to aligning their business and political strategies with globally agreed targets.
This includes the UN Kunming-Montreal Global Biodiversity Framework and Global Framework on Chemicals, which focus on protecting nature and tackling environmental harm caused by chemicals and waste.
The investors say companies must enhance their sharing of information on how their business impacts biodiversity as well as develop strategies to transition to making products that are safe and sustainable.
Failing to address chemical pollution exposes companies and their investors to financial risks, they argued.
The group highlighted how increased public awareness and scientific understanding of the long-term health and environmental consequences of chemicals has led to a rise in litigation and regulation.
The statement has been co-ordinated by responsible investment group ShareAction, Achmea Investment Management, ChemSec, Erste Asset Management, IEHN of Clean Production Action, Planet Tracker and Mercy Investment Services.
Signatories include BNP Paribas Asset Management, Rathbones Group, Caisse des Depots et Consignations, SVVK-ASIR, Swedbank Robur and Impax Asset Management.
It comes as Government officials, business leaders, scientists and campaigners gather in Uruguay this week for discussions on the next phase in the UN Global Framework on Chemicals.
Alexandra Pinzon, head of biodiversity at ShareAction, said: 'Chemical companies have a huge role to play in curbing pollution, which would help address the interlinked crises of biodiversity loss and climate change.
'The majority of manufactured products, from fertilisers and paints to make-up and clothes, rely on chemicals, but the toxicity and pollution associated with these chemicals is wreaking havoc on ecosystems and damaging human health.'
Julie Gorte, senior vice president for sustainable investing at Impax Asset Management, said: 'This statement on the importance of tackling pollution and biodiversity loss has never been more welcome.
'If we want to avoid a planetary catastrophe, we must all act to achieve a circular economy and eliminate the pollution that is one of the major drivers of biodiversity loss.'
Arthur van Mansvelt, senior engagement specialist at Achmea Investment Management, said: 'This statement shows investors are deeply concerned that the chemicals sector is not sufficiently mitigating the risks related to biodiversity loss from pollution.'
In a separate policy-focused statement, more than 40 investors with nearly four trillion dollars in assets under management also highlighted the crucial role that regulation plays in enabling the transition of the chemicals industry to safe and sustainable products.
The group outlined recommendations for governments around the world to strengthen and harmonise global policy frameworks on chemicals to support this transition.
Hashtags

Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles


The Guardian
5 hours ago
- The Guardian
Why we need a right not to be manipulated
Many nations already enshrine a right not to be defrauded, and even a right not to be deceived. If a company sells you a new medicine, falsely claiming that it prevents cancer, it can be punished. If a firm convinces you to buy a new smartphone, saying that it has state-of-the-art features when it doesn't, it will have violated the law. But in the current era, many companies are taking our time and money not by defrauding or deceiving us, but by practising the dark art of manipulation. They hide crucial terms in fine print. They automatically enrol you in a programme that costs money but does not benefit you at all. They make it easy for you to subscribe to a service, but extremely hard for you to cancel. They use 'drip pricing', by which they quote you an initial number, getting you to commit to the purchase, only to add a series of additional costs, knowing that once you've embarked on the process, you are likely just to say 'yeah, whatever'. In its worst forms, manipulation is theft. It takes people's resources and attention, and it does so without their consent. Manipulators are tricksters, and sometimes even magicians. They divert the eye and take advantage of people's weaknesses. Often they exploit simple ignorance. They fail to respect, and try to undermine, people's capacity to make reflective and deliberative choices. A manipulator might convince you to buy a useless health product, not by lying, but by appealing to your emotions, and by painting seductive pictures of how great you will feel once you use the product. Or they might tell you an anecdote about someone just like you, who used a supposed pain-relief product and felt better within 12 hours. Anecdotes have real power – but they can be profoundly misleading. More insidiously still, manipulators might know about, and enlist, some of the central findings in contemporary behavioural economics, the field that explores how people depart from perfect rationality. All of us are vulnerable in this regard, subject to the 'cognitive biases' elaborated by Daniel Kahneman, Amos Tversky, Richard Thaler and others, that affect our behaviour. These can be hard to recognize, and harder still to overcome. For example, human beings tend to suffer from 'present bias'. We care a lot about today and tomorrow, but the future is a foreign country, Laterland, and we are not sure we are ever going to visit. Tactics like 'buy now, pay later' take advantage of this. Another bias is 'loss aversion'; we tend to dislike losses a lot more than we like equivalent gains. That's why advertisers might claim 'you can't afford not to' buy their product. Inertia is a powerful force, and companies exploit 'status quo bias' by automatically subscribing you to something in the knowledge that even if it's possible to opt out, many won't bother. Our attention is limited, which means we are able to focus on only a subset of the things that come across our radars. Knowing this, manipulators present only the most attractive aspect of a transaction and downplay other, less inviting parts. So, manipulation is all around us, and rarely punished. But if we aim to create a right not to be manipulated, we will have to specify what we are talking about. A moral right can define manipulation broadly. A legal right should focus on the worst cases – the most egregious forms of trickery, those that are hardest to justify and that are most likely to impose real harm. Those worst cases occur when people are not given clarity that they are committing themselves to certain terms – and when the terms are ones they wouldn't consent to if they had full knowledge. For example, there should be a prohibition on billing people in accordance with terms to which they did not actively agree, unless it is clear that they would have agreed if they'd been asked. The underlying principle should be one of personal autonomy, which means that hidden fees and costs should be banned too. We know that rules designed to bring those fees and costs into the open can do a great deal of good. A couple of recent examples from the US: in 2024, the Department of Transportation created a rule that requires airlines and ticket agents to disclose charges for checked baggage, carry-on baggage, changing or cancelling a reservation and so on up front. Also in 2024, the Federal Communications Commission required internet service providers to display standardized 'broadband nutrition labels'. These include details of pricing, data allowances and broadband speeds, and enable customers to compare providers' offerings like-for-like, without tricks and obfuscation. But consumer protection is only the start. In 1890, two lawyers, Samuel Warren and Louis Brandeis, proposed a new right: the right to privacy. They were not entirely clear about its contents, but the bedrock was the 'right to be let alone'. Warren and Brandeis's thinking helped to launch a thousand ships, including rules against the disclosure of private facts, against surveillance, and around personal choices (including the right to same-sex marriage). The right not to be manipulated now is a lot like the right to privacy back in 1890. At this stage, we cannot identify the full scope, and the appropriate limits, of that new right. The protection of consumers and investors is urgent. How it might apply to politics is a more delicate matter, and lawmakers will need to tread cautiously there. One thing is clear, though: manipulation is a threat to our autonomy, our freedom and our wellbeing. We ought to be taking steps to fight back. Professor Cass R Sunstein is the co-author of Nudge and founder and director of the Program on Behavioral Economics and Public Policy at Harvard. His new book, Manipulation: What It Is, Why It's Bad, What to Do About It, will be published by Cambridge in August (£22). To support the Guardian order your copy at Delivery charges may apply. Thinking, Fast and Slow by Daniel Kahneman (Penguin, £14.99) Misbehaving by Richard H Thaler (Penguin, £10.99) The Psychology of Money Morgan Housel (Harriman House, £16.99)


Telegraph
6 hours ago
- Telegraph
Markets would be mad to play a game of chicken with Trump now
'This market just wants to go up.' I have heard this said in the past month on trading floors on both sides of the Atlantic. Market speculators tend to express this kind of sentiment when share price performance is strong but risks to the global economy seem to point to the downside. As major global stock markets break all-time highs, investors are fretting about what may happen on Aug 1 when Donald Trump reinstates the so-called reciprocal tariffs, which have been paused since April 9. Recall that the US president was forced to hit pause to arrest an unfolding financial panic. He sparked fears that the sky-high tariffs he had unveiled a week earlier would trigger an inflationary slump in the US and tip parts of the world economy into recession. In the week following the tariff announcement, the Dow Jones Global Index fell 11pc. Since then, however, stocks have enjoyed a precipitous rise. Now, unlike in April, when sudden tariffs jarred risk appetites, market participants are sanguine about their return. Some market makers talk of a Taco trade – an acronym for 'Trump always chickens out'. They bet that the president will not have the stomach to risk another painful market correction and will kick the can down the road again. After all, the initial 90-day pause should have ended on July 8 but was extended to Aug 1. Trump has built a reputation as a strongman leader who is willing to take big risks. Inviting him to play a game of chicken seems too much like an opportunity that he will welcome in order to prove his mettle once again. The reason for the market rally, more likely, is much more banal and relates to more fundamental developments. First, now that the initial surprise has worn off and thanks to the pauses, businesses and consumers have had time to adjust supply chains, shift inventories, and spread out the costs. Second, because the US has struck deals with a few major economies, the threat of a full-blown tit-for-tat global trade war has been contained. The UK and Japan agreements are a positive signal for the difficult ongoing negotiations with the EU, Canada, and Mexico, which together account for almost half of all US imports. Third, the all-important US–China talk s are running on a separate schedule. After the dramatic escalation in March and April, which amounted to a de facto trade embargo after the baseline US tariff on Chinese goods peaked at 145pc, both sides agreed a 90-day pause in Geneva on May 12. US tariffs are temporarily set at 55pc, while China applies tariffs of 10pc on US imports. Scott Bessent, the US treasury secretary has signalled that the deadline on the US–China talks will likely be extended beyond the Aug 12 deadline. The Yale Budget Lab calculated that the average effective US tariff was 22.5pc on April 7 – the highest since 1909. Once all current and likely deals are accounted for, the eventual average-weighted US tariff will still be the highest it has been in more than a century and little different from the one that Trump announced on 'liberation day'. In its latest update on July 23, which incorporates Trump's deal with Japan (this reduced the baseline tariff from a threatened 25pc to a still-high 15pc), the average weighted tariff was 20.2pc. But US imports are only worth about 3pc of global GDP, and Polymarket odds of a US recession in 2025 have dropped from 65pc in April to 20pc at present. With more time to adjust and with less uncertainty, the dislocations associated with US protectionism have become more manageable. Greater clarity over US trade policy has allowed a host of broader and less prominent positive themes to play a greater role in shaping expectations in financial markets. They include the fading Russian energy shock in Europe, the global demand boost coming from the lagged impact of interest rate cuts by major central banks, the massive defence-and-infrastructure-oriented fiscal expansions in Europe and China, and further tax cuts in the US. In a world where one or two big risks dominate headlines, individual investors can often struggle to see a way through the fog. But prices in markets, which reflect collective fundamentals and the implicit wisdom of the crowd, can tell a different story. Global equities consistently provide a reliable leading indicator for global trade and production. In uncertain times, the market is often the best economist. For instance, in late 2020, as major economies locked down for the upcoming Covid-19 winter wave and economists – including myself – warned that it could take the global economy years to recover from the unfolding mega-recession, global equities hitting all-time highs proved to be the signal for the V-shaped recovery that subsequently took place in 2021. Then, as now, the market's collective wisdom offers valuable clues for those willing to pay attention.


Geeky Gadgets
12 hours ago
- Geeky Gadgets
How to Build Reliable AI Agents : Create AI Systems That Never Fail
What if the AI agent you deployed today could not only predict user needs with precision but also adapt seamlessly to unforeseen challenges? The demand for reliable, scalable AI systems has never been higher, yet the path to building them remains fraught with complexity. Developers are inundated with a dizzying array of tools, frameworks, and trends, many of which promise innovation but deliver unpredictability. The stakes are high: a poorly designed AI agent can lead to skyrocketing operational costs, inconsistent outputs, and even reputational damage. But here's the good news—by focusing on core principles and strategic simplicity, you can cut through the noise and build AI agents that are not only functional but also dependable in real-world applications. In this step-by-step overview, Dave Ebbelaar shares actionable insights to help you master the art of creating robust AI agents. You'll discover how to strategically integrate Large Language Models (LLMs), design workflows that prioritize predictability, and implement recovery mechanisms that keep your system operational under pressure. From understanding the critical role of memory in maintaining context to using human-in-the-loop feedback for high-stakes tasks, this guide will equip you with the tools to navigate the challenges of AI development. Whether you're a seasoned developer or just beginning to explore the field, the principles outlined here will help you design systems that don't just meet the demands of 2025 but anticipate the needs of the future. After all, reliability isn't just a feature—it's the foundation of trust in AI. Building Reliable AI Agents Understanding the Challenges in AI Development The AI development landscape is increasingly crowded with new tools and frameworks, often accompanied by significant hype. This can make it challenging to identify practical solutions that balance innovation with reliability. Common challenges include high operational costs, unpredictable outputs, and overly complex workflows. To overcome these obstacles, it is essential to focus on foundational principles rather than fleeting trends. By prioritizing simplicity, deterministic engineering practices, and strategic use of LLMs, you can avoid these pitfalls and create systems that are both efficient and reliable. Core Principles for Building Reliable AI Agents To ensure your AI agents are effective and dependable, adhere to the following core principles: Avoid over-reliance on pre-built frameworks: Custom solutions tailored to your specific use case provide greater control, flexibility, and adaptability. Custom solutions tailored to your specific use case provide greater control, flexibility, and adaptability. Minimize LLM API calls: Use LLMs strategically to reduce operational costs and mitigate dependency risks, making sure your system remains efficient and cost-effective. Use LLMs strategically to reduce operational costs and mitigate dependency risks, making sure your system remains efficient and cost-effective. Adopt deterministic engineering practices: Design workflows that are predictable, modular, and easy to debug, enhancing the overall reliability of your system. These principles serve as the foundation for creating scalable and robust AI systems capable of meeting the demands of diverse applications. How to Build Reliable AI Agents Watch this video on YouTube. Master AI agents with the help of our in-depth articles and helpful guides. The Seven Foundational Building Blocks Building reliable AI agents involves integrating seven critical components into your workflows. Each component plays a unique role in making sure functionality, reliability, and scalability. 1. Intelligence Layer The intelligence layer forms the core of your AI system, handling reasoning and context-based tasks. By incorporating LLM API calls strategically, you can enable advanced language processing while maintaining simplicity and adaptability. This layer should be designed to evolve with changing requirements, making sure long-term flexibility and relevance. 2. Memory Memory is crucial for maintaining context across interactions. Whether managing conversation history dynamically or storing data in a database, memory ensures a seamless and coherent user experience. This is particularly important for applications such as virtual assistants or customer support systems, where long-term context retention is essential for effective communication. 3. Tools for External Integration External integration tools enable your AI agents to interact with APIs, databases, and other systems. These tools extend the functionality of LLMs beyond text generation, allowing your agents to retrieve information, execute commands, or update records in external systems. This capability is vital for creating versatile and functional AI solutions. 4. Validation Validation ensures that your AI agent produces structured and consistent outputs. For instance, using JSON schemas to validate data formats can help maintain quality assurance. This step is particularly important in production environments, where errors can have significant consequences and undermine system reliability. 5. Control Control mechanisms, such as routing logic and if-else statements, enable deterministic decision-making. By categorizing tasks and directing processes, you can modularize workflows and enhance system predictability. This approach is especially useful for managing complex operations efficiently and effectively. 6. Recovery Recovery mechanisms are essential for handling errors, API failures, and rate limits. By incorporating retry logic and fallback strategies, you can ensure that your AI agent remains operational even under adverse conditions. This resilience is critical for real-world applications where reliability is a non-negotiable requirement. 7. Feedback Feedback systems introduce human oversight into your workflows. For high-stakes or complex tasks, adding approval steps ensures that sensitive operations are reviewed and validated. This human-in-the-loop approach is particularly important for applications where errors are unacceptable and accountability is paramount. Key Insights for Developers When designing AI agents, it is important to distinguish between general-purpose assistants like ChatGPT and specialized, fully automated systems. Structured outputs and context engineering can significantly enhance reliability and performance. Additionally, debugging and logging are essential for understanding the decision-making processes of LLMs. For high-stakes tasks, incorporating human-in-the-loop systems provides an additional layer of oversight and accountability, making sure that your AI agents operate with precision and reliability. Practical Implementation Strategies To implement these building blocks effectively, focus on creating modular workflows that integrate the strengths of each component. Consider the following strategies: Use Python or similar programming languages to develop custom solutions for memory management, validation, and error handling. Prioritize error recovery and fallback mechanisms to ensure robustness in production environments. Use LLMs judiciously, reserving their use for tasks that truly require advanced language processing capabilities. This modular approach simplifies development while enhancing the scalability and adaptability of your AI systems, making them better suited to meet the demands of diverse applications. Building for the Future By mastering these foundational building blocks, you can design AI agents that are reliable, scalable, and adaptable to a wide range of use cases. Focus on first principles, break down complex problems into manageable components, and use LLMs strategically to maximize their impact. This approach will empower you to create robust AI systems that not only meet the challenges of 2025 but also remain relevant and effective in the years to come. Media Credit: Dave Ebbelaar Filed Under: AI, Technology News, Top News Latest Geeky Gadgets Deals Disclosure: Some of our articles include affiliate links. If you buy something through one of these links, Geeky Gadgets may earn an affiliate commission. Learn about our Disclosure Policy.