Dr Anandhi Vivek Dhukaram PhD MBA

Overview and basic terminology of LLM evaluations 🧑‍🏫 and benchmarking. Can recommend reading if you are new to these topics We will keep adding to this blog series as we move forward with the DevQualityEval 🏇

7

The UK Research and Innovation's new policy document on how to use AI tools responsibly is probably much more widely appropriate than the intended research funding application process. Some highlights: - Never input sensitive or personal information (without consent) - Consider the risk of bias - Ensure outputs are not falsified, fabricated, plagiarised, misrepresented I also find the assessor angle on not using AI tools to 'outsource' personal evaluation and judgement very relevant but likely hard to apply in practice in many scenarios. One of the core use cases at the moment for LLMs like ChatGPT is summarising meetings, long documents, etc. Without validating the accurateness and completeness of the summary, "AI judgement" is very quickly baked into any subsequent human decision making. #AI #Policy #Ethics #HumanFactors #DecisionMaking The full policy is here: https://lnkd.in/dDD6B3mC

3

1 Comment

Are we prepared to regulate AI effectively? Dive into this thought-provoking piece from Holly Anderson as she reviews respondents' replies to the government's AI White Paper consultation. https://riaabg.uk/4awt4qQ #ArtificialIntelligence #Regulation #TechEthics

1

So many different courses and consultancies available here at Howe2 Training & Consultancy 👍 Take a look at: https://lnkd.in/en5pE3dN If one you are looking for isn't listed, it is probably being built at the moment - let us know and we can make sure it is available for you 🙏 We also offer "bespoke" courses. This means your policies and procedures can be embedded into the course(s) at no financial to yourselves at all 🚫💷 #training #consultancy #reallifetraining #specialisttraining

2

It is still unclear what if any changes to the law and regulations will be required. It is an interesting area to look at control by laws versus the benefits of a more liberal approach.

1

The UK Government is advancing responsible AI use in public services through the new Algorithmic Transparency Recording Standard (ATRS). This initiative ensures transparency and accountability as AI accelerates decision-making, reduces backlogs, and improves citizen support, aligning with the “Plan for Change” to modernize services and drive economic growth. Departments like the Foreign Office and the Department for Business and Trade are already using AI for faster assistance and targeted business support. ATRS sets clear guidelines for transparency, detailing AI tools' data, technologies, risks, and the role of human oversight, fostering trust and ethical innovation. #AI #PublicSector #ResponsibleAI #DigitalTransformation #AlgorithmicTransparency #EthicalAI #DataPrivacy Article Link: https://lnkd.in/gbtSUNGe

1

These are subjects that are vital for the future of anyone in the technical content world.

9

1 Comment

Always wonderful to hear Christina Silver, PhD’s deep insights on how to navigate the world of qualitative data. We have for so long been on the sidelines but the AI explosion is propelling this research front and centre. Keep up to date and don’t forget the core statistic principals.. Leximancer #ai #qualitativeresearch #llm #rag #thematicanalysis #multivariantanalysis #hierarchialclustering #baysian #cassandra #lexi6

1

This OECD - OCDE report provides a definition of #AI incident and related terminology. These definitions aim to clarify what constitutes an AI incident, an AI hazard, and associated terminology without being overly prescriptive. This document will inform the development of a common AI incidents reporting framework and its application through the OECD AI Incidents Monitor (AIM).

10

Running a digital service business and preparing for #EAA regulations in 2025? Derisk your hashtag #AccessibilityAudit preparation with a free screen reader initial diagnostic by our team ✅ Our team has experience of visual challenges and blindness for a true user understanding 🧑 In-house technical and software team for actionable recommendations 🖥️ Limited space is available, so message us to claim your slot 🚀 #DigitalServices #UserExperience #Accessibility #EAA

3

Getting Knowledge Transfer (KT) correct can be a challenge for organisations who have deployed a consultancy to deliver complex transformation. Failure to execute an effective Knowledge Transfer workstream can result in a reliance to keep the consultancy onboard for longer than desired, hence driving up external post-implementation costs. Clients regularly voice frustrations at this, but the responsibility should be split two ways between the client and consultancy. Here are our top tips to ensure clients can contribute to effective Knowledge Transfer workstreams. ✍️ Let us know your thoughts in the comments! #KnowledgeTransfer #Consultancy #ProjectDelivery

7

UP NEXT: AI Regulation in the UK: Striking a Balance Between Innovation and Safety. This panel can delve into the UK's approach to AI regulation, including any emerging frameworks and how organisations can innovate while ensuring ethical and responsible AI use​. Join this session: https://lnkd.in/eh6CDPSR #RISK #PrivSec #AI #Safety #AIRegulation #EthicalAI #Ethics #Trust

10

🚥 Using an LLM based AI? Careful, it most likely is not EU AI compliant! A colleague in an AI Safety group here in the UK said he had been participating in a whole day OECD workshop on PETs (Privacy enhancing technologies), something I co-authored a paper on a few years ago. He said 'it threw up an interesting EUHC law case that would imply (and I asked the speaker to confirm my understanding is correct) that *most if not all LLMs/GenAI tools are currently non-compliant with EU law (where the privacy of individuals is concerned)!!!* He further commented within the backdrop of the EU AI Act, suggesting it's going to be "v.interesting" to see the implications of this particular case!' 🚩🚩These are flags I started waving in the August 2023 Digital Leaders conference, specifically on the dangers and dividends of using AI in Executive Coaching. 🏛I'll be talking about my EU AI Compliance for Coaching report tomorrow prior to meeting with UK Digital and Technology trade representatives discussing AI Safety with the UKs Secretary of State for Science and Innovation. I'll be discussing the ethics of coaches adapting to AI and adopting it with Dr Fabio Oliveira (PhD) and Dr Naeema Pasha from Henley Business School. Artificial Intelligence -As a Coach should you Adapt or Adopt? Tuesday 14th May 2024 11:00-12:30 London, UK Time Zoom Meeting. https://lnkd.in/ea-z3VDu

9

6 Comments

Having worked at over 10 research intensive organisations, I've seen many recurring challenges in research management. Surprisingly, despite variations in size, structure, and internal policies, most organisations face similar issues in relation to research project management. One persistent challenge across the board has been timesheets! Regardless of the organisation’s size large or small completing and signing off timesheets in a timely, accurate manner has remained a consistent struggle. Nearly two decades on, it’s surprising to see this issue still crop up, especially given that staff costs often represent over 70% of the awarded budget on any given project. With such a crucial component to research recovery rates and with staff costs relating to such a large proportion of the budget amount. Timesheets are crucial for accurate budgeting and project cost recovery. So, as an industry, can we afford to let another decade pass without addressing this? Especially when tools and systems exist to streamline this process and ensure they are completed accurately and on time. Why are timesheets still an issue? Let’s discuss what systems are teams using? If I have tagged you below I would love to hear how you manage timesheet completion in your team and if it works well or the process needs improving? #timesheets #research #postaward #ukri #H2020 #horizoneurope #innovateuk

1

1 Comment

An excellent support scheme to boost your data science teams #construction #transport #logistics #warehousing #agriculture #foodprocessing #creative industries #aiforbusiness #ai #data #productivity #uk #ukbusiness #skills #datascience

2

"The UK Government’s AI blueprint signals an exciting moment for innovation. Digital twins and data-sharing infrastructure are central to realising this vision, not just as tools for technological advancement but as catalysts for regional growth and resilience. By aligning AI Growth Zones with existing regional initiatives, we ensure that innovation is inclusive, impactful, and sustainable." Justin Anderson, Director, Digital Twin Hub The UK Government’s ambitious 50-point plan to ‘turbocharge #AI’ marks an exciting milestone for #innovation across the nation. The Digital Twin Hub at Connected Places Catapult welcomes this vision, which emphasises long-term investment in AI capabilities, infrastructure and collaboration. #Digitaltwins and data-sharing infrastructure are critical enablers, providing tools to model, test, and optimise real-world systems and amplify effectiveness of AI solutions. The plan’s focus on AI Growth Zones highlights the opportunity to align innovation with the #devolution agenda, creating a stronger foundation for regional growth and impact: 📌 AI Growth Zones: Regional innovation Establishment of AI Growth Zones is designed to boost AI innovation by creating hubs of expertise and investment. The zones can play a pivotal role supporting UK regional economies by aligning with and amplifying the impact of existing Innovation Zones. 📌 Building regional ecosystems through data-sharing AI Growth Zones can advance data-sharing infrastructure, creating localised networks to connect public, private, and academic stakeholders. These ecosystems are vital for enabling digital twin technologies, which require high-quality, interoperable data. 📌 Scan > Pilot > Scale: A proven path to regional success The Government approach to AI mirrors the successful methods employed by the Catapult Network, instrumental in fostering regional innovation. By combining this with the resources of AI Growth Zones, local authorities and devolved administrations can accelerate development and deployment of AI-driven solutions. 📌 Shaping data for #regionalgrowth Government emphasis on shaping data collection aligns perfectly with the goals of AI Growth Zones. By prioritising creation of new, high-value datasets to address local needs, these zones can empower regions to harness AI and digital twin technologies. 📌 Driving collaboration through Knowledge Hubs The AI Knowledge Hub complements AI Growth Zones by fostering #collaboration and knowledge-sharing across regions. These hubs ensure that innovation does not occur in silos but as part of a coordinated, nationwide effort. Read more: https://lnkd.in/dVC8D8uf Erika Lewis Dr Alison Vincent, CBE FREng Mark Enzer Anne-Marie Friel Sarah Cameron Peter van Manen Ben Ford Charles Kennelly Liz St Louis Simon Evans Davin C. Anthony Tuffour George Economides Mark Coates FCIHT, FCInstCES Cristina Data Matt E. David Wagg Melissa Zanocco OBE Natasha Good Department for Science, Innovation and Technology

44

2 Comments

⬇️ This is one to read ⬇️ As we gather in Brussels for #AIGG24 and discuss what the EU AI Act will mean in practice, this report lays out how those leading the charge are building and thinking about AI governance programs.

21

There is no innovation without collaboration. Innovation thrives when diverse stakeholders—industry, academia, the public sector, and startups—collaborate. Each brings unique strengths to the table, creating a synergy that propels groundbreaking advancements. Industry: Converts theoretical concepts into practical applications, leveraging resources and market access to scale innovations. Academia: Generates cutting-edge research, ensuring it has practical implications through partnerships with industry. Public Sector: Provides funding, regulatory support, and public-private partnerships, fostering an environment conducive to innovation. Startups: Bring agility and disruptive ideas, challenging norms and exploring new business models. When we collaborate we can facilitate: 1. Resource Sharing: Ensures projects have necessary funding, knowledge, and tools. 2. Diverse Perspectives: Fosters creative problem-solving. 3. Risk Mitigation: Spreads risk, enabling ambitious projects. 4. Accelerated Development: Reduces time from concept to market. 5. Broader Impact: Ensures innovations benefit various societal segments. Collaboration is essential for driving innovation. By leveraging each other’s strengths, these stakeholders create a vibrant ecosystem where groundbreaking ideas flourish, benefiting society as a whole. What are your favourite ecosystem examples? #collaboration #ecosystem

12