Start Up No.2150: DPD chatbot gets sweary, you are not an embassy, the Big IT failure, how journalists beat the Post Office, and more


The San Francisco city council has spent half a million dollars dithering over its replacement for a public rubbish bin. CC-licensed photo by Rafael Castillo on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 10 links for you. You say trash, we correctly say rubbish. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.


DPD AI chatbot swears, calls itself ‘useless’ and criticises delivery firm • The Guardian

Jane Clinton:

»

The delivery firm DPD has disabled part of its artificial intelligence (AI) powered online chatbot after a disgruntled customer was able to make it swear and criticise the company.

Musician Ashley Beauchamp, 30, was trying to track down a missing parcel but was having no joy in getting useful information from the chatbot. Fed up, he decided to have some fun instead and began to experiment to find out what the chatbot could do. Beauchamp said this was when the “chaos started”.

To begin with, he asked it to tell him a joke, but he soon progressed to getting the chatbot to write a poem criticising the company.

With a few more prompts the chatbot also swore.

Beauchamp shared the conversation on X, with the chatbot replying to one message: “Fuck yeah! I’ll do my best to be as helpful as possible, even if it means swearing.” Then in another instance, the chatbot calls itself a “useless Chatbot that can’t help you”.

One post by Beauchamp, a classical musician from London, was viewed 800,000 times in 24 hours. Referring to the chatbot, he wrote on X: “It’s utterly useless at answering any queries, and when asked, it happily produced a poem about how terrible they are as a company.”

DPD uses AI in its online chat to answer queries as well as human operators. The company said a new update had been behind the chatbot’s unusual behaviour and it had since disabled the part that was responsible and was updating its system as a consequence.

«

“A new update” suuuuuuuuure. Expect plenty more like this. First contact with the enemy: people primed to screw things up. But now let’s see what Sam Altman thinks.
unique link to this extract


ChatGPT is best for people in these industries: OpenAI CEO Sam Altman • CNBC

Tom Huddleston Jr.:

»

Anyone can ask ChatGPT to answer a question or perform a task. But the popular chatbot is particularly useful for workers in three specific industries, according to OpenAI CEO Sam Altman.

“Coding is probably the single area from a productivity gain we’re most excited about today. It’s massively deployed and at scaled usage, at this point,” Altman said during a recent episode of “Unconfuse Me,” a podcast hosted by Bill Gates. “Healthcare and education are two things that are coming up that curve that we’re very excited about, too.”

Altman, whose company makes ChatGPT, made a point of noting that today’s AI systems “certainly can’t do [those] jobs” for you. But in those three fields, workers might benefit from using the chatbot as a productivity tool, he said.

«

Strange how he doesn’t mention marketing spiels and advertising and particularly helpline chatbots, but maybe that would sound a bit trivial.
unique link to this extract


You are not an embassy • How To Survive The Internet

Jamie Bartlett:

»

One of the trickier aspects of digital life is the constant pressure to opine. To have a strong opinion on a subject, and to share it with the world. It’s literally baked into the design of the most popular platforms. ‘What’s happening!?’ barks X. (Interestingly, until April 2023, X merely asked ‘What is happening?’ Notice the subtle difference of tone.) ‘What’s on your mind, Jamie?’ wonders Facebook. Some of the finest minds in the world work extremely hard to encourage you to tell everyone what you’re thinking and feeling. No wonder it’s hard to resist.  

Well – what is happening? (?!) If you watch the news, lots of bad things. And so we become something of a one-person embassy, pronouncing publicly where we stand on the key moral issues of the day. Anything from a government announcement, to a think-tank report, to a tragic event. We are ‘Devastated to hear about….’ We are ‘Outraged by the news of… ‘ And we are ‘Disappointed to see that [insert name] supports…’   

If I am honest, I know very little about most bad things going on in the world. Certainly not enough that sharing my view will inform or educate or enlighten. Yet whenever I see a news report, an urgent need rises up: what shall I say about this? I have a feeling about it – which must be shared! (And ideally in emotionally charged language, since that will receive more interactions). 

What’s wrong with calling out the bad stuff going on? Nothing per se. And certainly not on an individual level. The problem is when people feel a soft and gentle pressure to denounce, to praise, to comment on things they don’t feel they fully understand. Things they don’t feel comfortable speaking about. Things that are contentious and difficult to discuss on heartless, unforgiving platforms where the wrong phrase or tone might land you in hot water.  

What social media has done is to make silence an active – rather than the default – choice. To speak publicly is now so easy that not doing it kind-of-implies you don’t know or don’t care about what’s going on in the world. Who wants to look ignorant or indifferent? And besides, who doesn’t want to appear kind or wise, or morally upstanding in front of others?  

But the result is an undirected anger from all sides: frenetic, purposeless, habitual and above all moralising.

«

“You are not an embassy” is such a good point. I’m always amazed by people who think that their holding an opinion will somehow make a difference. Bartlett is always worth reading.
unique link to this extract


No more ‘Big IT’: the failed 90s model has ruined too many lives • FT

Mike Bracken:

»

Since the 1990s, both Labour and Conservative ministers have handed control to “the Big IT crowd”. Successive prime ministers bought the line that major-project databases and applications would deliver great outcomes. They didn’t.

This orthodoxy still prevails. Entire services — tax, immigration, passports, benefits — have been given over to IT suppliers and their favoured consultancies, who in the absence of true competition can inflate contract values in return for maintenance and minimal changes. The resulting pattern is services becoming fixed, more expensive to run and unable to adapt.

Meanwhile, organisations born in the Internet era, from payments to retail sectors, have demonstrated how to deliver services cheaply, at global scale. Gareth Davies, the National Audit Office chief executive recently recommended the UK adopts “manageable projects compared to gigantic, overambitious attempts to change the whole world with one IT system.” He estimates we could save £20bn. I think that’s modest.

We already know what to do because it has been tried with success elsewhere. Namely design services with a “test and learn” approach that adapts as user needs and behaviour change, rather than making one big bet on an IT tender that tries to predict requirements years in advance. Inside government, put experienced operators rather than generalist policymakers in charge. And crucially, reform the Treasury approach: selecting single IT suppliers on long-term contracts creates the legacy IT arrangements that bedevil the public sector.

Horizon is just another painful chapter in the long story of Big IT failure: no empathy for users, leaders that do not understand the value of service delivery and technology, and a civil service culture that refuses to put procurement, commercial and technology skills on a par with policy.

«

Here’s something from Bracken’s post last week that’s worth considering:

»

In 2013-14, 70 nurses at the Princess of Wales Hospital were suspended as they were suspected of falsifying blood glucose results from patients. Five were charged by police, two went to prison. An expert witness uncovered that the digital records had been affected by the WannaCry cyber attack, and an engineer from the system supplier had accidentally deleted some of the records while making repairs. The IT team at the hospital was aware that this had happened, but the investigation team was not, and proceeded on the basis that computer evidence is always correct. It wasn’t. More ruined lives.

«

unique link to this extract


Our tiny magazine exposed the Post Office. They underestimated us • The Sunday Times

Katie Gatens:

»

The ITV drama Mr Bates vs The Post Office brought the struggle of the postmasters into the nation’s living rooms and back on to the floor of the Commons, resulting in legislation that should overturn all the wrongful convictions in one go.

But it was [Rebecca] Thomson, now 41, and Computer Weekly’s chief reporter, Karl Flinders, 51, who were the first to really take on the story. In 2010, Thomson left the title and Flinders took over reporting. He has since written 350 articles for Computer Weekly on the scandal. Theirs is the story of how a tiny trade publication of ten editorial staff exposed the biggest miscarriage of justice in the UK –– and persevered for 15 years.

Thomson tracked down and spoke to seven postmasters for her initial article, including Bates and Castleton. She also found Jo Hamilton, who was spared prison after villagers raised the money for her, and Noel Thomas, who was sentenced to nine months in jail and spent his 60th birthday behind bars.

But to her disappointment her investigation, published in 2009, didn’t have the desired effect. “I thought there’d be a much bigger reaction,” she says. “I thought people would look at it and say, clearly questions need to be answered.”

Thousands of viewers were left distraught by the trauma portrayed by actors and Thomson says speaking to the real victims was emotionally draining.

“It’s really upsetting to hear someone say, ‘I’ve been diagnosed with a terrible illness and still they won’t give me any compensation’,” she says. “You start to feel guilty for putting them through it, especially if you feel like you’re not getting the results that you wish you were getting.”

Computer Weekly had a small team of lawyers who they called on for that first article, but much of their reporting had to be scaled back for fear of being taken to court. The team were careful to strike out mentions of the Japanese tech firm Fujitsu, which created the software that led to the errors. “We had to be careful in that first piece, because we thought they might sue,” says Thomson. “It was hard to stay motivated especially because there were so many setbacks.”

«

A long time ago I worked at Computer Weekly; the team competed furiously with two other weekly computer trade publications for all sorts of scoops and exclusives and investigations. The teams were bigger; but also national publications took more notice of what we wrote. Though as Flinders explains, the Post Office told reporters on nationals the stories were rubbish, putting them off the deep investigation required for this story.
unique link to this extract


Twin Pics: create this image with AI

»

Use AI to create a matching image. Your goal is to match the image below as closely as possible. The closer you get, the higher your score.

«

Some of the people are getting 90%+, though quite how it’s measured might be open to question given that you can get 50% for something you wouldn’t really call “a match”.
unique link to this extract


Microsoft ‘senior leadership’ emails accessed by Russian SolarWinds hackers • The Verge

Tom Warren:

»

Microsoft has discovered a nation-state attack on its corporate systems from the same Russian state-sponsored group of hackers responsible for the sophisticated SolarWinds attack. Microsoft says the hackers, known as Nobelium, were able to access email accounts of some members of its senior leadership team late last year.

“Beginning in late November 2023, the threat actor used a password spray attack to compromise a legacy non-production test tenant account and gain a foothold, and then used the account’s permissions to access a very small percentage of Microsoft corporate email accounts, including members of our senior leadership team and employees in our cybersecurity, legal, and other functions, and exfiltrated some emails and attached documents,” says the Microsoft Security Response Center in a blog post filed late on Friday.

Microsoft says the group was “initially targeting email accounts” for information about themselves, but it’s not clear what other emails and documents have been stolen in the process. Microsoft only discovered the attack last week on January 12th, and the company hasn’t disclosed how long the attackers were able to access its systems.

“The attack was not the result of a vulnerability in Microsoft products or services. To date, there is no evidence that the threat actor had any access to customer environments, production systems, source code, or AI systems,” says Microsoft.

The attack took place just days after Microsoft announced its plan to overhaul its software security following major Azure cloud attacks.

«

As to the timing of the blogpost, it seems to have been published around 9pm GMT, 4pm EST, 1pm PST. There’s certainly a faint element of “taking out the trash” – publishing something near the end of the business week so that it gets minimal pickup. Sorry if we spoiled that, Microsoft.
unique link to this extract


SF’s long, expensive odyssey to build bespoke trash cans stalls • San Francisco Chronicle

Aldo Toledo:

»

San Francisco Public Works has paused its quest to deploy new bespoke trash cans across the city amid a looming budget deficit, the department told the Chronicle on Friday.

The department, which has spent more than half a million dollars on the project, said in a statement Friday that it’s moving forward with the new trash can — the Slim Silhouette design, which won in a three-way contest in 2022 — but that “the procurement may be put on hold because of the city’s significant budget shortfall projections.”

The statement says that the city is “in the midst of the budget process” and that Mayor London Breed has asked all departments to look for significant savings of at least 10%. “At this point, everything is on the table,” Public Works spokesperson Beth Rubenstein said, adding that the agency has held off on issuing a request for proposals “because of the projected budget shortfall and the current budget process.”

The update from Public Works comes nearly a year after the department announced it had decided on the Slim Silhouette to replace aging green trash cans the city has been using since 1993. Those trash cans have long been criticized for often being soiled, filled to the brim with garbage or, paradoxically, making street corners dirtier. 

To replace the cans, San Francisco spent an eye-popping $537,000 to test three prototypes. The cost was later lowered to $400,000, and after many tests, a final prototype was selected. The city has still spent more than $500,000 on the entire effort, though updated costs were not immediately available. 

In summer 2022, the city deployed six custom trash can models across the city, giving residents a chance to weigh in on their favorite design. 

Aside from a slight hiccup in May when officials in the Civic Design Review Committee put the process on hold because of skepticism about the new design’s effectiveness, Public Works has been mum on its plans for finding a manufacturer and rolling out the sleek, silver cans. The department estimated in 2022 that the 3,000 cans would be ready in 2023.

«

Astonishing. The bins are too full so we need to redesign them? Why not more emptying runs?
unique link to this extract


National Grid: Live

»

Great Britain’s exposed position in the north-east Atlantic makes it one of the best locations in the world for wind power, and the shallow waters of the North Sea host several of the world’s largest offshore wind farms.

New wind power records are set regularly, and between 9:00am and 9:30am on 21st December 2023 British wind farms averaged a record 21.81GW of generation.

…An open source project by Kate Morley.

…The data comes from National Grid ESO’s Data Portal, Elexon’s Balancing Mechanism Reporting Service, and the Carbon Intensity API (a project by National Grid ESO and the University Of Oxford Department Of Computer Science).

«

Given we’ve just experienced one of the strongest winter storms, that record might have been broken.
unique link to this extract


Lots of good news – and good numbers – again in offshore wind • WindEurope

»

Things are looking up again for offshore wind in Europe. 2023 saw a record 4.2GW of new offshore wind farms come online, up 40% on 2022. And €30bn of new investments were confirmed – covering 9GW that’ll be built over the coming years. The supply chain is also seeing a turnaround, with new factories announced in Poland, Denmark, Germany, the Netherlands and Spain.

2023 was the best year on record for new offshore wind installations in both the EU and across Europe as a whole. Europe built 4.2GW of offshore wind in 2023. That’s 1.7GW more than in 2022. Of that, 3GW was in the EU, an increase of 2.1GW year on year.

The Netherlands, France and the UK installed the most new capacity. This includes the 1.5 GW “Hollandse Kust Zuid” project in the Netherlands – now the world’s largest operational wind farm.

Offshore wind investments in Europe also reached a new record. A total of €30bn was raised across 8 wind farms. This will finance 9 GW of new offshore capacity. This record comes after legal uncertainty and unhelpful market intervention had led to a drop in offshore wind investments, falling to an all-time low of €0.4bn in 2022. It also means that projects which had to postpone their final investment decision in 2022 are now moving ahead – excellent news for Europe’s energy security and competitiveness.

«

Would not have guessed a couple of decades ago that wind would play such a gigantic part in renewables. But here we are.
unique link to this extract


• Why do social networks drive us a little mad?
• Why does angry content seem to dominate what we see?
• How much of a role do algorithms play in affecting what we see and do online?
• What can we do about it?
• Did Facebook have any inkling of what was coming in Myanmar in 2016?

Read Social Warming, my latest book, and find answers – and more.


Errata, corrigenda and ai no corrida: none notified

3 thoughts on “Start Up No.2150: DPD chatbot gets sweary, you are not an embassy, the Big IT failure, how journalists beat the Post Office, and more

    • Wellllll, yes, often, but we’re also familiar with the Ambassador for Whereveristan coming on the media to explain that actually, the ruler isn’t a dictator AT ALL, dearie me no.

  1. I’m not amazed by “opinion will somehow make a difference”, as there’s a large amount of “glurge” in general about how you, yes YOU, can make a difference in the world – simply by speaking out and standing up for justice. Social Media is one particular case, but it’s hardly confined to there. Plenty of discussion shows will have someone pushing that idea.

    The reality is actually pretty depressing: No, generally, you can’t make a difference. The fact is that the world is shaped by disputes among billionaires, politicians, celebrities, etc, and their servants, and you are none of that. Definitely nobody outside your immediate family and friends cares what you think, and likely they don’t care either. Don’t bother if it’s anything other than wasting time.

    That’s not something which will garner an invitation to do a TED talk.

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.