News
Researchers develop artificial intelligence that can identify cancer cells
Scientists from Osaka University in Japan have developed artificial intelligence (AI) that can identify different types of cancers based on microscopy images of their cells. The AI was also able to determine whether the cancer cells were resistant to radiation, and further learned the differences between human and animal cancers. Since the accuracy and timeliness of traditional methods of identifying cancer cells are prone to delays and errors, an accurate and automated system for accomplishing this would be beneficial to cancer research and treatment overall. The results of the scientists’ research were published in the December 2018 issue of Cancer Research.
The type of AI developed for the cell identification is called a convolutional neural network (CNN); it’s loosely based on the connectivity patterns used by neurons in the brain and primarily used for classifying images. As described in their publication, the scientists used a training set of 10,000 images each of human cervical cancer cells (ME-180) and mouse squamous cancer cells (NR-S1), e.g., the thin, flat types of cells found on the surface of the skin and thin linings around various organs throughout the body. They also included images of radioresistant clones with the set, and ultimately obtained a 96% accuracy in a 2,000 image test.
Because the types of cells in a single cancerous tumor can vary widely, identifying the specific cells present is important for determining the best treatment. Thus, having a tool to provide this information quickly and accurately could have a significant impact. The Osaka team hopes to expand the types of cancers their AI can identify and ultimately establish a universal system that can identify all cancer cell types.
Using artificial intelligence in the battle against cancer is being explored throughout the world as the number of uses devised expands. In one notable instance, a team of scientists from The Institute of Cancer Research in London and the University of Edinburgh has developed an AI technique called REVOLVER (repeated evolution of cancer) which identifies DNA mutation patterns in cancers to predict the ways they will change in response to treatment. Similar to how bacteria become resistant to antibiotics, so too can cancers become resistant to the drugs used against them. By removing the unpredictability variable in cancer behavior, scientists would be able to stay ahead of the disease’s progress and tailor treatments accordingly.
The collaboration between AI and healthcare overall is growing, not just in cancer research – even Google is making contributions to the field. Fortunately, the agencies regulating developments are also attuned to the changes. Earlier this year at the AcademyHealth 2018 Health Datapalooza, the U.S. Food and Drug Administration (FDA) Commissioner Scott Gottlieb, MD signaled the agency’s positive position towards the field. “AI holds enormous promise for the future of medicine, and we’re actively developing a new regulatory framework to promote innovation in this space and support the use of AI-based technologies,” he stated at the event. He also referred to the agency’s plans to streamline their regulations and tools to be sufficiently flexible to handle the rapid pace of advancements and “focus on the ways in which real-world data flows.”
A new report from Reuters claims that a transport authority in Sweden is pushing back against the approval of Tesla’s Full Self-Driving suite because it will travel over speed limits.
The report says the Swedish Transport Administration (TRV) recommends the European Union votes against FSD’s approval. TRV believes it should not be approved until Tesla disables FSD’s ability to speed.
TRV sent a letter to the European Union’s Technical Committee on Motor Vehicles (TCMV), which is set to meet on June 30 to discuss the potential approval of the Tesla FSD suite in the country. Tesla, which has received various approvals in Europe over the past two months, has not provided a comment.
Teslas operating on FSD do travel over the speed limit, depending on the Speed Profile that is chosen. Drivers have the ability to disengage FSD at any point; Tesla specifically states that those supervising the suite are responsible for its actions.
Let’s cut to the chase: humans operating any vehicle speed almost daily in the United States. Realistically, speed limits in the U.S. are more frequently treated as speed minimums. However, other countries are different, and driving behaviors are less aggressive.
TRV believes that “allowing automated systems to systematically exceed legal speed limits…risks undermining both the legal framework and the expected safety benefits of vehicle automation,” the report stated. It’s surprising that Tesla has not received this claim from other countries previously.
This could be a good argument to bring Max Speed back, the setting that previously allowed the driver to choose the absolute fastest the car would travel.
This would still put the responsibility of supervision in the hands of the driver. It would allow the driver to choose whether the car would travel over the speed limit or not, acknowledging that they set the speed, and if they get pulled over, there would be no ability to argue it.
However, it does not seem as if this is something Tesla will do, especially considering many U.S. drivers have requested the feature in an effort to eliminate speeding or at least tone it down. The company has not shown any interest in bringing it back.
Tesla has approvals for FSD in Europe in Estonia, Lithuania, Denmark, the Netherlands, and Belgium.
Tesla is going to let you guide Full Self-Driving with Grok in 3 months, CEO Elon Musk confirmed on X.
The response from Musk, which revealed Tesla plans to allow drivers to effectively control the car and its navigation more explicitly using Grok, puts the feature for about September.
A Tesla owner said that Full Self-Driving is great, but owners should be able to “converse with Grok like we can with an Uber driver.” She then used examples like, “Grok, turn right here,” and “Drop us off right here, we’ll walk due to traffic,” and finally,” Drop at entrance first, then park far away.”
Coincidentally, the final piece of dialogue would also mean features like Banish are potentially on the way soon.
This functionality will be there in about 3 months or so
— Elon Musk (@elonmusk) June 18, 2026
Banish is also referred to as “Reverse Summon,” and would enable the car to self-park while dropping occupants off at their destination.
This would be a great way to improve the overall experience while supervising FSD. Navigation is already a major painpoint that many owners complain about. Manual overrides when a maneuver is requested or canceled (like using the turn signal stalk to override a navigation route), do not always work.
The feature could be especially useful in street parking scenarios in a city, where spots are sometimes tough to come by. Many of us who grab dinner in a more populated area will park a street or two over from wherever we’re going, because sometimes you know that’s the best you will get. If a driver using FSD could say, “Hey Grok, turn right here on Queen St. and park in that open spot on the right,” it could save a lot of confusion FSD might have on its own.
Musk teased that a similar feature was “coming” back in February:
Tesla Full Self-Driving set to get an awesome new feature, Elon Musk says
It is certainly surprising that Tesla is doing it at this point. The company’s more recent moves have been more evident of taking control and inputs away from humans and putting them in the AI’s hands more frequently. The biggest example of this was taking away Max Speed in AI4 cars, giving us Speed Profiles, and not having any input on the fastest speed the car will travel.
Of course, giving navigation preferences to Grok is availble already in Teslas, but not at the drop of a hat. Instead, you can suggest a certain route at the beginning of your drive.
Here’s an example of that from December:
🚨🏈 I am taking my parents and Fiancee to the @Ravens game next weekend and asked @Grok to help me route my @Tesla through a specific neighborhood to reach the correct Lot we will park in.
This is a great example of the new @grok nav integration with the Tesla Holiday Update: pic.twitter.com/rPp4I7q8Yv
— TESLARATI (@Teslarati) December 13, 2025
Finally, the original post that Musk responded to mentioned a parking preference after dropping off the occupants, which describes the Banish feature that Tesla has teased for years.
We’re not sure if Musk was responding more to the ability to guide the car with Grok, or whether he also was including Banish in the three-month prediction timeframe.
A close-up image of a Cybercab engineering vehicle in Peabody, Massachusetts, reveals a compact triangular side repeater camera housing equipped with an integrated washer mechanism.
This seemingly small hardware addition could prove to be one of the most critical components for achieving reliable, unsupervised Full Self-Driving (FSD) — not just for the dedicated Robotaxi but potentially for existing AI4-equipped vehicles as well.
The washer system’s importance cannot be overstated in Tesla’s vision-only autonomy approach. Cameras are the sole sensory input for the neural networks powering FSD, constantly interpreting the environment for safe navigation. In real-world conditions, however, lenses quickly accumulate rain, snow, mud, dust, or road spray.
Many of us Tesla owners, especially those who deal with any sort of winter weather at all, know the all-too-common alert that pops up when cameras are obstructed:
Even brief obstructions can drop perception confidence, trigger safety disengagements, or force the vehicle to pull over, although these are relatively rare. Instead, most of the time, the camera will need a wipe from the owner next time they stop the car.
But unlike human drivers who can manually clear their view, a Robotaxi operating 24/7 without a steering wheel or mirrors must maintain pristine vision autonomously. The Cybercab’s side repeater washer delivers targeted cleaning bursts precisely where needed for merging, lane changes, and blind-spot monitoring — functions that demand uninterrupted visibility from the external cameras:
And this is how the side camera and washer look like on a Cybercab. This is from an Engineering vehicle in Peabody MA. pic.twitter.com/Re8VknpmLM
— Tobias Goebel (Unsupervised) (@tpgoebel) June 17, 2026
This hardware directly tackles a known pain point in current FSD deployments. Owners frequently report camera-related alerts during inclement weather, which is understandable, but needs to be solved for a true autonomous experience.
For a production Robotaxi fleet aiming for high utilization and minimal downtime, robust washer systems represent a foundational reliability upgrade; essentially, they’re a must-have. Early sightings suggest the design may extend to rear cameras as well, creating a comprehensive cleaning architecture that keeps the entire vision suite operational in harsh environments.
Without it, even the most advanced neural nets struggle when their “eyes” are compromised.
What Does This Mean for AI4 Cars?
This Cybercab detail raises timely questions for AI4 cars already on the road. While Hardware 4 delivers superior compute and camera resolution compared to earlier versions, production models typically lack dedicated side and rear washers. Tesla has included them on Model Y robotaxis that it is using in the fleet:
Tesla Robotaxi has a highly-requested hardware feature not available on typical Model Ys
As Tesla refines unsupervised FSD for broader release, the gap in environmental resilience becomes evident. Software improvements can help mitigate issues, but they cannot fully replace physical cleaning in heavy rain or muddy conditions. Analysts and owners increasingly speculate that AI4 vehicles may eventually require similar washer retrofits — or a future AI4.5 variant — to match the Cybercab’s all-weather readiness and support the same level of autonomy.
As testing progresses, the Cybercab’s washer mechanism highlights Tesla’s pragmatic focus on real-world robustness. It may well become the hardware piece that determines how quickly and reliably FSD scales from prototypes to everyday vehicles.
Tesla faces Full Self-Driving pushback in EU over ‘speeding’
Tesla teases greater Grok FSD integration and ‘Banish’ feature ‘in about 3 months’
