How AI can support peacebuilding efforts

Hi!

Today's key developments in AI highlight its potential role in sustainable peacebuilding, boosting job performance for less experienced workers, aiding space exploration, and the increasing wariness of Japanese firms towards its use in job applications. The optimistic view of AI sees it bringing in comprehensive understanding, inclusivity, and legal handling capacities to human-led efforts. A study by MIT Sloan indicates a significant productivity boost for less experienced workers, while NASA leverages AI for real-time issue detection and resolution in space missions. But it's not all plain sailing as Japanese companies express concern about students' use of AI in job applications, making personality assessment challenging. In the AI tools sector, Perplexity AI emerges as an accurate conversational search engine. Lastly, the VC scene buzzes with Databricks acquiring AI startup MosaicML for $1.3 billion and Thomson Reuters purchasing AI legal tech startup Casetext for $650 million, signaling noteworthy movements in the AI M&A landscape.

Slicing into chucks:

• ✌🏼 How AI can support peacebuilding efforts

• 👩🏻‍💻 Workers with less experience gain the most from generative AI

• 🚀 NASA is planning to help astronauts using AI-powered assistant

• 🇯🇵 Japanese firms wary of students' use of AI in job applications

✌🏼 How AI can support peacebuilding efforts

AI has the potential to support human efforts in achieving sustainable peace by mitigating obstacles and improving the chances of peace. Despite concerns about the misuse of technology and the existential risks of AI, an optimistic view reveals how AI can help address major challenges such as violent conflict. However, the approach will have to overcome inherent issues of AI like biases and the risk of imposing a Western-centric view of conflict resolution. AI should be used to support, not replace, humans in conflict resolution strategies, ensuring a diverse range of perspectives. AI could assist in human-led peacebuilding efforts by addressing issues more effectively in areas where humans fall short, like ensuring a comprehensive understanding of a conflict’s historical and cultural environment, inclusivity in dialogue, and handling legal complexities. Large Language Models (LLMs), a subset of generative AI, can process information that surpasses human intelligence, providing a holistic, up-to-date understanding of the conflict at the negotiating table, improving inclusivity in negotiation, and enhancing early warning systems. AI technologies can also be harnessed for long-term peacebuilding, offering comprehensive legal analysis, supporting accountability mechanisms, and fostering an inclusive, innovative, and democratic society.

👩🏻‍💻 Workers with less experience gain the most from generative AI

A recent study suggests that generative AI significantly benefits workers with less experience, providing a fast-track improvement in their job performance. The study, conducted by associate professor Danielle Li at MIT Sloan, PhD candidate Lindsey Raymond, and Stanford University professor Erik Brynjolfsson, observed that contact center agents who had access to a conversational AI assistant experienced a 14% boost in productivity, with the most significant gains affecting inexperienced or low-skilled workers. Generative AI, especially large language models, are capable of inferring relationships between inputs and outputs, and are not just reliant on explicit instructions. In a practical context, at a customer contact center of a Fortune 500 company, the use of generative AI led to an increase in the number of customer chats resolved per hour by 13.8%. Workers using the AI model managed to resolve 2.5 chats per hour within two months, compared with 1.7 chats per hour for those not using the model, who needed eight months to hit the higher threshold. The study also highlighted that the use of AI led to improved customer sentiment, with fewer requests to speak to a manager and quicker transfers to other departments.

🚀 NASA is planning to help astronauts using AI-powered assistant

NASA is reportedly developing an AI assistant similar to ChatGPT to assist astronauts in the field, allowing users to communicate with the spacecraft directly. The AI system is designed to detect and alert mission operators of potential glitches and inefficiencies in real-time, negating the need for engineers to be sent into space whenever a space vehicle goes offline or encounters software issues. This AI tool is expected to be deployed on the Lunar Gateway, a planned space station, without reliance on supercomputers. An alternative approach called federated learning is being explored, allowing a fleet of rovers to share knowledge without immediately sending data back to Earth, thereby conducting distributed learning collaboratively. While this approach reduces the amount of data sent to Earth, it may still necessitate a local data storage system on the space station.

🇯🇵 Japanese firms wary of students' use of AI in job applications

A Jiji Press survey found that Japanese companies are apprehensive about students using generative AI in their job applications. The survey, which included about 60 major domestic companies and received responses from 38, revealed concerns about accurately understanding students' personalities from their AI-generated employment applications. Nine of the 38 responding companies expressed a "generally negative" viewpoint on the use of generative AI in job applications. For instance, East Japan Railway Co. (JR East) expressed a desire for students to articulate their thoughts about the company in their own words. Similarly, convenience store operator Lawson Inc. stated its preference to evaluate applicants' unique characteristics and qualities. An official from a manufacturing company argued that making employment decisions based on materials that do not reflect students' thoughts wouldn't benefit either the companies or the applicants.

🛠️ AI tools update

Perplexity AI is a conversational search engine developed by a team of AI experts from OpenAI, Meta, Quora, and Databrick

Perplexity AI is designed to be more accurate than traditional search engines and can also act as a chatbot. When a user inputs a question, the model scours the internet to give an answer and displays the source of the information it provides. Perplexity AI can be accessed for free from their website and is extremely easy to use. Users can ask questions, find step-by-step instructions, define terms, find opening hours, and find the weather forecast. Perplexity AI has raised $26 million in series A funding in March 2023.

💵 Venture Capital updates

Databricks strikes $1.3 billion deal for generative AI startup MosaicML

Databricks, a firm that sells software tools for building AI systems and advocates for open-source models, has agreed to acquire artificial intelligence startup MosaicML in a deal valued at $1.3 billion. This acquisition aligns with the current dominance of AI in the tech industry, with significant investments flowing into AI startups. MosaicML, based in San Francisco, offers software tools designed to make AI work more affordable, typically involving the training of AI algorithms on large data sets with costly computer chips. The deal will merge Databricks' AI technology with MosaicML's language-model platform, providing businesses with a simple and cost-effective way to maintain control, security, and ownership of their data. Both companies have released open-source foundation models, a core technology behind services like OpenAI's ChatGPT. Since its founding in 2021, MosaicML has raised $64 million from investors, including Lux Capital and DCVC.

Thomson Reuters buys Casetext, an AI legal tech startup, for $650M in cash

Thomson Reuters has agreed to acquire Casetext, a Y Combinator-backed legal tech startup, for $650 million in cash. The deal is expected to close in the second half of 2023, pending regulatory approvals. Casetext, founded in 2013, has evolved from a platform for attorneys to share knowledge and access legal texts, to a provider of AI-driven legal tools and workflows. Its primary product, CoCounsel, leverages AI to review documents, prepare depositions, and analyze contracts, with backing from OpenAI’s GPT-4 language model. Casetext boasts a customer base of over 10,000 law firms and corporate legal departments, and has raised more than $64 million in funding.

🫡 Meme of the day

⭐️ Generative AI image of the day

Before you go, check out how HR leaders should be cautious of these costly A.I. risks.