Thomas Brewster

🚨 Alert: Foreign Espionage Targeting British Academic Systems 🚨 In a recent briefing with vice-chancellors from 24 top universities, MI5 sounded the alarm on the escalating threat of foreign states targeting British academia. The aim? Undermining national security by exploiting cutting-edge research for their own military and economic gains. See: https://lnkd.in/epWwQRPH This revelation underscores the critical need for heightened vigilance and protective measures within our academic institutions. With no specific state named, concerns persist regarding undue influence, particularly from China, as highlighted by Parliament's intelligence and security committee. Drawing parallels with recent events in Canada, where Chinese espionage has targeted intellectual property and interfered with federal elections, it's evident that this isn't an isolated issue. The access of Chinese agents to advanced technology and research in Canada has raised significant concerns for national security and economic prosperity. The implications are dire: intellectual property theft, influence on research agendas, and potential misuse of partnerships. The Canadian government's response to such threats underscores the seriousness of the issue, as they seek to safeguard their own breakthroughs in fields like artificial intelligence. As we contemplate measures to fortify our academic research from foreign interference, it's imperative to strike a balance between innovation and security. Academic institutions are businesses and they make a lot of money from foreign students, but at what risk. Is there enough being done to vet students from potentially dangerous foreign states that are looking to use academia as cover to steal intellectual property or advance their own military and economic goals? For more on the operations of the MSS and MI5 goto https://lnkd.in/eBTeUtXi Transparency in funding, enhanced personnel vetting, and bolstered research security capabilities are among the strategies being considered. Deputy Prime Minister Oliver Dowden rightly emphasizes the importance of adapting to evolving threats while preserving the integrity of our esteemed institutions. The Russell Group and Universities UK, representing leading universities, reaffirm their commitment to national security and collaboration with government and intelligence agencies. We must remain vigilant and proactive in defending our academic assets against hostile actors. Together, we can safeguard our research excellence and uphold the values of academic freedom and integrity. #NationalSecurity #AcademicResearch #ForeignEspionage #MI5Warning

10

Big tech and social media companies have come under fire for their failure to adequately address the risks posed by disinformation and artificial intelligence (AI) deepfakes to the integrity of elections in the UK. The Joint Committee on the National Security Strategy (JCNSS) said big tech companies had responded to the committee’s inquiries on the risk to democracy posed by AI deep fakes and disinformation with “opaque and diffuse” statements that made it difficult for them to be held to account. #deepfakes #AI #elections #disinformation #misinformation #bigtech

13

1 Comment

I would recommend reading this call to action from from David Lacey - former Foreign Office CISO and consultant - if you are an IT security professional interested in joining like-minded people to discuss the application of Zero Trust in IT security.

7

1 Comment

Data profiling by political parties is mostly guesstimation of who you are. Machine-born caricatures guiding an industry that's capable of political interference, eventually looping back into voter behaviour. But democracy means putting power in the hands of the people ✊ Data protection gives us the right to find out what information political parties hold about us; what profile they've built about us. And we can opt-out of them using our data. Read more in our latest blog ⬇️ #GeneralElection #DigitalRights #DigitalRightsAreHumanRights #privacy #dataprotection #GDPR #ukpolitics https://lnkd.in/eR_eMiYp

9

AI is developing very quickly, which means the governance around its use will have to be ready to change as gaps occur. That emerges as the key message of a report published just before Parliament ended from the Science, Innovation and Technology Committee. One disturbing detail is the suggestion that the AI Safety Institute - a key feature of the existing governance arrangements - has not been able to get access to some developers’ models to carry out pre-deployment safety testing. It suggests there is already a tension between commercial interests and the effort to provide safety in how AI is used. https://lnkd.in/er9-BXxC #ai

9

Belatedly reposting this from Mi3Australia. Back on the podcast this week to discuss the ACCC "data firms" report, why it ended up being called that, the ever-closer competition + privacy law overlap and how even the privacy-review changes expected in July might not be enough, according to the ACCC. Thanks for having me, Paul McIntyre!

10

East Riding Council has adopted a formal policy for the use of generative AI. I wouldn't swear to it, but think this is a first for local government in the UK. Given the increasing use of the technology, and a sense that it could provide big gains in productivity, I should we'll see a lot more local authorities laying down clear rules for its deployment. While the policy lays out steps and safeguards that will be familiar to people in digital roles, it should provide useful guidance to those who are not familiar with debates on the safe and ethical use of digital tech. https://lnkd.in/eUkbQrGh #localgovernment #genai #generativeai

6

🌐A recent TechCrunch article by Natasha Lomas highlights the European Union's vigilant monitoring of AI technologies in the wake of the tragic FICO shooting and ongoing concerns over disinformation under the Digital Services Act (DSA). The EU's attention on X underscores the imperative for responsible AI deployment. The article sheds light on how policymakers are navigating the complex landscape of AI regulation amid heightened public scrutiny. DSA's Role in Addressing Disinformation: The DSA's ongoing probe into disinformation is a critical step toward combatting online misinformation and safeguarding digital spaces. These developments highlight the importance of proactive AI governance to foster trust and accountability. Our organization remains committed to advancing ethical AI practices and welcomes constructive dialogues on the responsible use of AI technologies. For more insights, read the full TechCrunch article here: https://lnkd.in/eZPA-RhK Let's continue shaping a future where AI serves humanity responsibly and ethically. #AIethics #Regulation #TechPolicy #DigitalServicesAct #EthicalAI

2

1 Comment

What are your thoughts on the UK’s role as a global leader in AI safety with the introduction of the Inspect tool? How do you think such initiatives will influence the future of AI development and ethics worldwide? Please follow AI Workplace and let us know in the comments! #artificialintelligence #ai #airegulation #aisafety #responsibleai #aiethics

4

A politics of fairness: Transformation in how we live must come, one way or another https://lnkd.in/eDKN_SZm

1

1 Comment

"VERDICT: An affecting book and always slightly to the side, as indeed were all of Kafka’s writings." Starred review (!) from Library Journal for LIFE AFTER KAFKA, by Magdaléna Platzová, out August 6 in my translation from Bellevue Literary Press: https://bit.ly/3Vh1tF4

15

For the first time, two Pulitzer winners disclosed using AI in their reporting. Andrew Deck's article (link in comments) is such an interesting read. He describes how these two winners used innovative approaches. 1. "City Bureau and Invisible Institute trained a custom machine learning tool to comb through thousands of police misconduct files." They asked 200 volunteers from the community to annotate these files to train the model. Trina Reynolds-Tyler "hopes other reporters who rely on machine learning tools understand the value of embedding in the community they are reporting on, and grounding their data work in real places and real people. 'We must make it our business to bring people to the future with us.'" 2. "The New York Times visual investigations desk trained a model to identify 2,000-pound bomb craters in areas marked as safe for civilians in Gaza. They trained an object detection algorithm on Picterra to identify these craters. "We didn’t use AI to replace what would’ve otherwise been done manually. We used AI precisely because it was the type of task that would’ve taken so long to do manually that [it would distract from] other investigative work," explained Ishaan Jhaveri. It's worth noting that these two stories were based on machine learning models for specific tasks, not the big shiny chatbots. --- Also worth reading: The NYT is assembling a team of five people dedicated to AI initiatives, an interesting mix of editorial, engineering, and design. The team includes Juliana Castro Varón, Rubina Fillion, Dylan Freedman, Duy Nguyen, and Zach Seward. “We will always use A.I. as a tool in service of our mission; transparently and ethically; and with human guidance and review,” state their guiding principles.

120

7 Comments

New review: Duncan Redford on David Kenyon's _Arctic Convoys: Bletchley Park and the War for the Seas_. London: Yale University Press, 2023. https://lnkd.in/e-4NsScU "_Arctic Convoys: Bletchley Park and the War for the Seas_ by David Kenyon is an excellent and long overdue book. It explores the role of signals intelligence at the operational and tactical levels on the Arctic Convoys of the Second World War and is firmly based on the rigorous analysis of primary sources, some of which have not been exploited. As Kenyon rightly points out, the general view of the Arctic Convoys was determined by the popular understanding of the Convoy PQ-17 catastrophe. In 1942, Convoy PQ-17, the first significant joint Anglo-American naval operation under British command, was ordered to scatter, resulting in the sinking of 24 out of the 35 Russia-bound merchant ships by German submarine and air attack. This book is therefore a useful corrective to the notion that the Arctic Convoys were only defined by PQ-17, and, more importantly, explores what was known, assumed, and feared at the time—not just for PQ-17 but for the other convoys as well...."

1

🚨 Time is running out for regulators to tackle the threats #AI poses to the democratic process ahead of July's election, warns the Alan Turing Institute. While there's no clear evidence that deepfakes have altered election results, early signs show AI-generated content could harm democracy. Recently, deepfake audio of Sadiq Khan and President Biden caused alarm. Sam Stockwell of The Alan Turing Institute highlights the difficulty of distinguishing between AI-generated and authentic content, which can be exploited to undermine democracy. The report calls for Ofcom and the Electoral Commission to issue guidance on the fair use of AI in campaigns and recommends media guidance and voter education on spotting AI-generated content. Regulators must act swiftly to ensure our elections remain resilient and voters maintain trust in the democratic process. 🌐🗳️ #AI #Elections #Democracy #Regulation #Deepfakes https://lnkd.in/eFGqhx2i

2

*** Today's AI Policy Daily highlights: 1. AI-generated deepfakes and misinformation are emerging as a major challenge ahead of elections, with concerns about phony robocalls and forged photos. 2. The UK government announced a new law targeting the creation of AI-generated sexually explicit deepfake images, with penalties of fines and possible jail time. 3. Microsoft is investing $1.5 billion in the Abu Dhabi-based AI group G42, as part of a collaboration between the Biden administration and Microsoft to limit China's influence in the Gulf region. 4. A bipartisan group of US Senators, including Mitt Romney, proposed a federal framework for AI oversight to mitigate national security risks. 5. The US government's over-reliance on Microsoft is a growing problem, as the company faces a series of cybersecurity failures. Here is today's edition: April 17, 2024 Check it out - click here: https://lnkd.in/e7etKdK3 #ai #artificialintelligence #aipolicy #aiprogramming #airegulation #aisafety 

1

Could disinformation disrupt the UK general election? I spoke to experts about the UK's approach to AI and the legal gaps it's created for potential disinformation to creep in ahead of next month's general election https://lnkd.in/ehuGziPZ

5

What do you think of AI Steve, the UK's first electoral candidate? https://lnkd.in/gA87SCmd Hoping to represent "the ultimate form of democracy," AI Steve claims to put policy power directly in the hands of constituents in the area of Brighton, UK, where it will stand for election alongside human counterparts. But does AI Steve really represent democracy in action? Can AI be trusted to act impartially? What kind of transparency is required to convince everyday people to trust such a thing to represent them? We'd love to hear your thoughts! gravity9 is not politically affiliated, nor related to AI Steve (however, we find the technology fascinating!) #AI #ArtificialIntelligence #AISteve

10

2 Comments

Some great points here. I thought 'limitless' was also a really good way of framing the usage for AI systems. If the technology starts to extend into every area of our lives, how can we begin to even imagine ways to ensure its safety? I was at a conference last week which focused on the use of AI in policing, and it was clear the Met Police throughly tested the algorithm underlying its live facial recognition technology - relying on independent advice from the 'National Physical Laboratory'. I'd never even heard of the organisation before so I went and looked it up. Apparently, it is a public corporation owned by the Department for Science, Technology & Innovation (DSIT) and its work is focused on 'developing and maintaining the national primary measurement standards', or 'metrology' - the science of weights and measures. I'm sure that was the right approach in this context, but I guess my point is do we have full confidence that the regulatory landscape for AI as it stands is clearly defined - and whose expertise should even be sought in the first place to ensure systems are safe?

1