Slashdot Mirror


40% of 'AI Startups' in Europe Don't Actually Use AI, Claims Report (forbes.com)

Two-fifths of Europe's AI startups do not use any AI programs in their products, according to a report that highlights the hype around the technology. From a report: Out of 2,830 startups in Europe who were classified as being AI companies, only 1,580 accurately fit that description according to the eye-opening stat on page 99 of a new report from MMC, a London-based venture capital firm. The label, which refers to computer systems that can perform tasks normally requiring human intelligence, was simply wrong.

"We looked at every company, their materials, their product, the website, and product documents," says David Kelnar, head of research for MMC which has $400 million under management and a portfolio of 34 companies. "In 40% of cases we could find no mention of evidence of AI." In such cases, he added, "companies that people assume and think are AI companies are probably not."

67 of 112 comments (clear)

  1. Coincidentally by Luthair · · Score: 5, Insightful

    No one does, at best people are using machine learning.

    1. Re:Coincidentally by Anonymous Coward · · Score: 5, Funny

      Reminds me of a screenshot I saw some days ago:
      "If it's written in Python, it's probably machine learning. If it's written in Powerpoint, it's probably AI"

    2. Re:Coincidentally by clevelandguru · · Score: 3, Informative

      AI is a broader term that includes Machine Learning, Natural Language Processing, Expert System... So even if a company uses Machine Learning, they are using an AI component.

    3. Re:Coincidentally by Ed_1024 · · Score: 1

      Exactly what I thought when reading the headline. AGI still seems to be a long way off...

    4. Re:Coincidentally by sfcat · · Score: 4, Insightful

      Reminds me of a screenshot I saw some days ago: "If it's written in Python, it's probably machine learning. If it's written in Powerpoint, it's probably AI"

      That quote is wrong...it should be, if its in Python, its applied math but is being called machine learning, if its in LISP or Prolog then its probably AI. Actually, no self-respecting CS person uses Python. Python is used by academics from other fields like Physics who then use their field's variant of mathematical analysis and call it AI. Then they turn down a person with an actual CS degree and experience in ML for a data science job because they don't understand some weird jargon that was invented two weeks ago. This is no surprise to me...the entire "Big Data" thing has been a fraud since the beginning and since the bosses have no idea what AI is, they can't tell when they are being conned. And apparently neither can the investors.

      --
      "Those that start by burning books, will end by burning men."
    5. Re:Coincidentally by Luthair · · Score: 1, Troll

      If someone says they have a rabbit, and they pull out a rabbit's foot would you agree with them?

    6. Re: Coincidentally by SharpFang · · Score: 1

      Would it put you in a position where you can write something that can be sold as AI though?

      Python gets the work done. I'd like to see a household appliance running on Prolog though.

      --
      45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
    7. Re:Coincidentally by phantomfive · · Score: 5, Insightful

      AI used to mean "machines that think like a human" (or at least close enough to be in Searle's Chinese room). After a while, AI turned out to be harder than everyone thought, so some people redefined AI to mean, "stuff we found while looking for AI." The original meaning of AI got renamed to "strong AI." That's in academia.

      In the outside world, the press and the media and popular culture, AI retains its original meaning: "machines that think like a human." Which is confusing for a lot of people when they hear scientists talking about AI, because strong AI has not been invented, and we have no idea even how to do it. We're missing some fairly important pieces.

      --
      "First they came for the slanderers and i said nothing."
    8. Re:Coincidentally by K.+S.+Kyosuke · · Score: 1

      That may very well be true, but there doesn't seem to have been a requirement for the "G".

      --
      Ezekiel 23:20
    9. Re:Coincidentally by Fnkmaster · · Score: 2

      At the same time, AI has come to be used in the tech press and mainstream press to mean "machine learning and related statistical techniques". Obviously this hurts the brains of many of us who still understand AI to mean "strong AI" or the newer moniker, "AGI" (artificial general intelligence), but we sort of have to roll with the language on this one.

      Wikipedia insists that machine learning is a subset of AI. OK, sure, I guess that's fine, in that it is one of a series of techniques that can provide reasonable performance in solving certain human-level intelligence tasks.

      When I'm speaking with a technical audience I go with more precise terminology - deep learning, reinforcement learning, unsupervised learning and try to avoid the headache, since we all know we are talking about statistical techniques in machine learning and none of it yet comes close to strong AI. Getting the mainstream press to correct their use of the abbreviation AI is about as likely as getting them to correct their use of the word "hacking".

    10. Re:Coincidentally by hazardPPP · · Score: 1

      if its in LISP or Prolog then its probably AI.

      What makes Prolog so special? It's a language based on formal logic, and formal logic is only one aspect of intelligence.

    11. Re:Coincidentally by phantomfive · · Score: 4, Informative

      AI has come to be used in the tech press and mainstream press to mean "machine learning and related statistical techniques".

      It really doesn't. Reporters mostly don't understand the difference between "machine learning" and "strong AI." They also don't understand the difference between "machine learning" and "statistical techniques."

      Read articles carefully: often the researcher/company will be saying one thing, and the reporter will hear "strong AI." Then we get that followed with "AI is a danger to humanity." In its current form it's not, and it will take a lot before AI is anywhere near capable of overtaking humanity. And sometimes the researchers encourage the confusion.

      --
      "First they came for the slanderers and i said nothing."
    12. Re: Coincidentally by phantomfive · · Score: 1

      How pattern recognition is not part of AI?

      It's most likely part of the solution to "how the human brain works," but it is far from the entire solution of general AI. That is why.

      --
      "First they came for the slanderers and i said nothing."
    13. Re:Coincidentally by eepok · · Score: 2

      Yes and no. It is USED as a broader term to mean all those things. However, it is deceptive to do so. Marketing departments try to distort existing vocabulary all the time with the intent of attaching current understanding to their very different product.

      Artificial Intelligence is one example. The use of the term "ridesharing" by app-based taxi services is another. "Bikeshare and carshare" are short-term bike and car rentals. In fact, the entire "sharing" economy is actually just "short term rentals" or "services provided by an independent contractor".

      It's all there to make you think it's something grander, innovative, and more meaningful than it actually is.

    14. Re: Coincidentally by Fnkmaster · · Score: 2

      I didn't say that it does mean that. I said it has come to be used that way. I support some prescriptiveness in language, but I am a realist about technical terms getting broad to the point of meaninglessness when they enter the mainstream. Fighting this is a frustrating and likely pointless endeavour (see: hacker).

      Your theory seems to be that writers at TechCrunch think when they say a startup is using "AI" to improve recruiting or the sales process or even to process large volumes of textual data that they really think they mean artificial general intelligence. I believe they are using a vague catchphrase or abbreviation that gets clicks, and that they know that's not "real AI" but don't care.

      Maybe the journalists at the New York Times or Washington Post are a more credulous lot and they really do think that is what it means, but I doubt it. They are all just capitalizing on confusion between Hollywood AI and machine learning for clickbait. That is what gets them paid.

    15. Re: Coincidentally by Luthair · · Score: 2

      the point
      your head

    16. Re:Coincidentally by skaralic · · Score: 1

      I just pronounce it "AL". I think it's more meaningful that way.

    17. Re:Coincidentally by Anonymous Coward · · Score: 1

      No, AI is a term that is broadly misapplied to include ML, NLP, ES, etc. Currently, there is no AI.

    18. Re: Coincidentally by phantomfive · · Score: 2

      Do you doubt it or not? You don't need to doubt, you can actually look at the articles and see how it's being used. Here's an example from tech crunch, that has diagnosed a computer as a psychopath. Whether the author believed it or not, that is how the term AI was used: https://techcrunch.com/2018/06...

      --
      "First they came for the slanderers and i said nothing."
    19. Re:Coincidentally by athmanb · · Score: 1, Interesting

      It's also because AI as a concept keeps moving its goalposts as we reach some of them.

      If you'd ask a guy from the 80s whether a device qualifies as AI that allows you speak the name of a destination, then leads you there on the optimal way using voice output, he'd say "of course yes". Nowadays we'd just say "durr no that's just Google Maps".

    20. Re: Coincidentally by phantomfive · · Score: 2

      You just made that up, you are imagining things. If you actually looked to see what people in the 80s were thinking about AI you almost certainly would have come across Searle's Chinese Room paper. Next time do a Google search.

      --
      "First they came for the slanderers and i said nothing."
    21. Re:Coincidentally by gweihir · · Score: 1

      Indeed. The whole term "AI" is a marketing lie. It has gotten to prevalent that researchers are now using "AGI" (Artificial General Intelligence), which essentially means the same thing but has not yet gotten corrupted by the marketing scum.

      Incidentally, there is no "learning" in machine learning either. It is just calculating parameters of statistical classifiers from data-sets. Calling that "learning" is about as wrong and dishonest as calling statistical classificators "AI". "Learning" and "AI" both imply insight. Insight is something machines cannot do and may well never be able to do.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    22. Re: Coincidentally by No+Longer+an+AC · · Score: 1

      Since I had to google it and I still have the tab open:

      The Chinese Room Argument

    23. Re:Coincidentally by LynnwoodRooster · · Score: 1

      Paul Simon thanks you!

      --
      Browsing at +1 - no ACs, I ignore their posts. So refreshing!
    24. Re: Coincidentally by arglebargle_xiv · · Score: 1

      In other news, 70% of blockchain startups don't use blockchain, except as a means of getting VC funding. Which is what the AI startups are doing too. And if you really want a way of getting VCs to bukkake money at you, do an AI blockchain startup.

    25. Re: Coincidentally by SharpFang · · Score: 1

      Still, you need *something* to show the VCs to attract them. Preferably something more interactive than PowerPoint presentation. So... Python for the rescue!

      --
      45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
    26. Re: Coincidentally by reanjr · · Score: 1

      Well, at least some of those machines are intelligent enough to know ML is AI. That makes them smarter than you.

  2. and the other 60% by Anonymous Coward · · Score: 1

    also donâ(TM)t use AI

    1. Re:and the other 60% by HornWumpus · · Score: 1

      They were doing god's work then.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
  3. Only 40%? by shatteredsilicon · · Score: 2

    Simple rule of thumb to distinguish machine learning from AI: - If it's written in Python, it's probably ML. - If it's written in Powerpoint, it's probably AI.

  4. No kidding by Brett+Buck · · Score: 4, Insightful

    Claiming that you are an "AI" company is just marketing buzz for most. Last year they would have been "blockchain startups", in 1999 they would have been "internet companies", and in 1960 they would have been somehow worked the word "rocket" into the name. It's been this way forever.

    1. Re:No kidding by DontBeAMoran · · Score: 1
      --
      #DeleteFacebook
    2. Re:No kidding by DontBeAMoran · · Score: 1

      And now we sit back and wait for "AI-powered" crawlers to automatically register this stupid-as-fuck domain name.

      --
      #DeleteFacebook
    3. Re:No kidding by Anonymous Coward · · Score: 1

      Yep, but you know how it is.
      Investor have read about it in some magazine so if you don't claim that your new line of teacup holders uses AI or blockchains or whatever the new buzzword is they will think that you are out of touch with the market and withhold the funding.
      Some of them are a bit behind and wants you to do the IoT thingy.

    4. Re:No kidding by LynnwoodRooster · · Score: 1

      This year at CES, down in the Eureka area, was a guy sitting in a booth. The curtain behind him had just 3 words, in relatively (1" high) small print..

      blockchain
      ai
      crypto

      His booth was actually quite busy. I take this as proof that, if you reach a critical mass of buzzwords, it does not matter WHAT you do, you WILL get funded as Homer Simpson found out...

      --
      Browsing at +1 - no ACs, I ignore their posts. So refreshing!
    5. Re:No kidding by prefec2 · · Score: 1

      It is not the companies claiming to be AI companies. It is analysts (other analysts) to claim companies do AI without evidence (claim by this analysts). So in short: Analysts claim other analysts do not work properly. What else is new?

    6. Re:No kidding by DarkOx · · Score: 2

      I don't even blame the companies. I blame the politicos and their moronic desire to constantly attempt to interfere in the market place. Cleveland for example has "block chain incubator."

      So if you are starting a business that does anything with a computer at all; you want to get free money and paraded in from of potential investors well you better find away to include block chain somehow even if that isn't the right tool for the job and if you can;'t manage that you need to work into the marketing literature how using blockchain as part of your product is an aspirational goal that you will realize "anytime now" because that is how you get the free handouts. Heck I would do the same!
       

      --
      Repeal the 17th Amendment TODAY! Also Please Read http://www.gnu.org/philosophy/right-to-read.html
    7. Re:No kidding by BarneyGuarder · · Score: 1

      eMachineLearningCloudBlockchainAI-as-a-service.com: We're democratizing social networks for the internet. Our current valuation is $4.5 Billion and our server closet is where the waiters keep their coats.

    8. Re:No kidding by The-Ixian · · Score: 1

      Next thing you are going to tell me is that my yoga pants aren't smart.....

      --
      My eyes reflect the stars and a smile lights up my face.
  5. It's probably even worse than that by locketine · · Score: 1

    I suspect MMC, a venture capital firm, doesn't have the expertise or resources to inspect the code, so how would they know that there's something like AI or ML powering the product? The article makes it sound like they simply looked at the company's marketing materials and maybe got a product demo from the marketing team.

    --
    Think globally but act within local variable scope.
  6. Analysts claim analysts are wrong by prefec2 · · Score: 4, Informative

    Analysts classify start ups to be AI companies while the analysts of the study claim that analysis is wrong, as they could not find evidence of AI used by the companies. The latter analysts also do not provide evidence on what data they come to this conclusion. All analysts do not provide any definition what this AI is and what they counts as AI in their respective studies.

    They do not analyze anything properly, they guess, use smoke and cloaks to confuse the audience.

  7. AI has now taken the place of blockchains by ChoGGi · · Score: 1

    for the sweet VC dollars.

  8. Re:New investment opportunity by SharpFang · · Score: 1

    Nobody. The acronym sucks.

    --
    45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
  9. Re:A case of cultural difference more likely by prefec2 · · Score: 4, Insightful

    The article is about analysts which claim other analysts made a wrong assessment based on data both do not have.

    So it is all bullshit. BTW they do not define what they mean by AI.

  10. What is AI? by zmooc · · Score: 2

    This all comes down to the definition of AI; these days, people tend to think the term is synonymous with machine learning, but it isn't nor has it ever been; AI is just about a system that's somehow "artificially intelligent". Rule-based systems and statistical analysis can be AI just as well. Us computer people that understand that simply following some rules is not "intelligent" may think otherwise, but the vast majority of humanity considers these things "intelligent". Any automated system that appears "intelligent" to its users or buyers is AI. By definition.

    --
    0x or or snor perron?!
    1. Re:What is AI? by Tony+Isaac · · Score: 1

      No. Just because the marketing people have led non-technical folks to believe that rule-based systems are "intelligent"...doesn't make it so.

  11. Oh, they're using is allright.,. by BarneyGuarder · · Score: 1

    AI right there in the name, so it is doing its job! What else is there? Also, 40% sounds low. Maybe it is higher over here in the states.

  12. This Is A Shocker by WankerWeasel · · Score: 1

    You're telling me there are companies capitalizing on our obsession with a current buzzword?

    1. Re:This Is A Shocker by iggymanz · · Score: 1

      let me check my Big Data for answer...

    2. Re: This Is A Shocker by iggymanz · · Score: 1

      I'll need some Data Scientists with Data Analytics to check their Dashboards

  13. AAI by goombah99 · · Score: 1

    Artificial Artificial Intelligence. It's the new genuine simulated leather.

    --
    Some drink at the fountain of knowledge. Others just gargle.
  14. Re:40%? It's 100%. by K.+S.+Kyosuke · · Score: 1

    but AI would indicate some notion of sentience

    No, it wouldn't. It has never been the requirement.

    --
    Ezekiel 23:20
  15. Re:New investment opportunity by hazardPPP · · Score: 1

    I'd love to invest in BAISCMDFISCMC! If they will accept East German marks that is.

  16. Fuzzy Logic by Anonymous Coward · · Score: 1

    I'm waiting for Fuzzy Logic powered Artificial Intelligence on a Transputer, then we'll be back in 1980.

  17. What does this remioned me of? by 3seas · · Score: 1

    It reminds me of the Website bubble where startup got investors to put money into startups that had no product or service.

  18. AAI Startups? by AdamStarks · · Score: 1

    Artificial AI Startups?

  19. Re: Missing the point. by Anonymous Coward · · Score: 1

    You don't know "neural nets" but you know "neural networks"? Do you also not know a cock up your ass but you know an erect penis in there?

  20. More proof so-called 'AI' is mostly marketing hype by Rick+Schumann · · Score: 1

    ..and now we see: most of what keeps getting hyped as 'AI' is marketing hype, just like I've been saying. Do you really want to trust your life to half-assed half-baked pieces of software?

  21. Re: Missing the point. by Anonymous Coward · · Score: 1

    We don't have a cogent definition of I, so AI is even harder to define, apart from the A part. Most of the time, though, people just want stuff that gets the job done cheaply, whatever it might be called, but verifying it is doing the job you think it is doing is not always easy.

  22. Most startups probably don't even use the "I". by mark-t · · Score: 1

    So saying that they don't use the artificial one either isn't news.

    And by the way... although this is just a personal pet peeve here, it's actually written as "A.I.", not "AI". It is an abbreviation, not an acronym.

    1. Re:Most startups probably don't even use the "I". by gweihir · · Score: 1

      Most startups probably don't even use the "I".

      So saying that they don't use the artificial one either isn't news.

      Oh, yes. Recently did a consulting job for an "AI" startup. What we should have told them is "Your idea is shit and anything solid you actually have is 20 years old." They had no clue about the state-of-the-art and were bumbling about incompetently. Surprisingly they had investors.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    2. Re:Most startups probably don't even use the "I". by HiThere · · Score: 1

      And do you also write "N.A.S.A."?

      --

      I think we've pushed this "anyone can grow up to be president" thing too far.
  23. Re:40%? It's 100%. by gweihir · · Score: 1

    The marketing scum cannot help themselves. They will corrupt any and all meaning of language if it means they potentially get one more sale.

    --
    Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
  24. Linear regression by lorinc · · Score: 1

    My personal experience with companies that claim to do AI, is that the vast majority of them claim it's AI as soon as they do a linear regression. Which always makes me wonder what they did before they switched to AI.

  25. Re:A case of cultural difference more likely by DCFusor · · Score: 1

    You win the internet today. About time someone called BS on the entire subject area. Artificial Ignorance, except nope, it's real.

    --
    Why guess when you can know? Measure!
  26. Re:More proof so-called 'AI' is mostly marketing h by jrumney · · Score: 1

    This seems like the opposite to me. If the claim is true, and a full 60% of AI startups are actually doing AI, then it would seem that there is more to it than the usual hype bubble. But probably the authors used an extremely lenient definition of AI, like if tensorflow is linked in somewhere, it is AI.

  27. Lots of reasons for stuff like this by reanjr · · Score: 1

    My employer has AI initiatives, but we just engaged in some creative explanations to convince zoning board we are an R&D company so we can move into our new office park.

    There are lots of reasons the reporting on these kinds of metrics is fuzzy.