Since providing my older answer to this question, and since considering electric scooters, I have actually purchased an electric bike and have to not update my old answer but provide a new much better answer with completely different perspective.
I used to dismiss electric bikes based on three problems:
- Bikes are low-speed high-torque devices whereas electric motors are high-speed low-torque devices. I used to think there's no logical place to put an electric motor into a bike. Motor in front wheel hub makes it hard to fix a front puncture, offers no logical way to implement reduction gearing for high-speed low-torque motors, makes it hard to have two wheelsets with different types of tires, and results in front wheelspin and a crash if riding on slippery steep uphill. Motor in rear wheel hub makes it hard to fix a rear puncture, has to be crammed in very small space because the rear hub also has bicycle gearing, offers no logical way to implement reduction gearing for high-speed low-torque motors, and makes it hard to have two wheelsets with different types of tires. Also any motor in hub requires a very high torque low speed motor which are uncommon beasts. Motor in bottom bracket would be even more ridiculous than motor in a wheel hub because the torque there needs to be much higher and the speed much lower, or so I thought.
- Law restricts electric bikes to 250 watts. This is not much higher than the power of your average cyclist when riding up a hill (actually athletic cyclists produce far bigger power levels).
- Law restricts electric bikes to 25 km/h. I thought this means that at 25 km/h the power assist needs to be already zero, so to avoid a very bumpy on-off-on-off-on-off assist, the power level needs to be reduced at already 22 km/h, or so I thought.
However, later I bought a new e-bike and found how e-bikes solve these challenges.
- Good quality e-bikes are mid-drives where the motor is at the bottom bracket. However, the bottom bracket area is made much larger and the motor is actually not directly connected to the bottom bracket, but rather connected via a reduction gearing. Thus you can have a high-speed low-torque motor and modify its speed and torque characteristics to be useful in bicycles.
- E-bike makers interpret the 250 watt limitation as "maximum sustained power". So it's very typical to have 70 newton meter e-bikes that provide the power even at 90 RPM. A simple calculation shows such a motor provides 659.73 watts, much higher than 250 watts. How can this be the case? The reason is that the e-bike has software that integrates the average power over a very long time period of an hour or so, and if it threatens to be the case that average power exceeds 250 watts calculated over a very long time, it starts to restrict power. Because an average cyclist rides on flat land at speeds higher than 25 km/h, on typical e-bikes the electric assist is used only when riding against a very strong headwind, up a hill or when accelerating. It is not a flatland assist level. Thus, almost never the 250 watt average power limitation is exceeded even though the motor assists at over 650 watts at every uphill.
- E-bike makers interpret the 25 km/h as the midpoint of where power begins to be reduced. So it doesn't mean that at 25.1 km/h the power needs to be already at zero. It's more like power begins to be reduced at 23.5 km/h and at 26.5 km/h it's at zero. So substantial power assist can be still provided at slightly above 25 km/h. Furthermore, because the electric motor in e-bikes is an uphill and acceleration assist device, even if some brand of e-bike already has the assist at zero at 25 km/h, it's not a major problem. Your average speed suffers but not by much.
So why do I advocate e-bikes over e-scooters for example? The reason is simple: range.
My e-road-bike has 125 - 175 km range using a half-kilowatt-hour battery, at the maximum assist level (125 km is the range on routes where I need to brake a lot, but a leisurely country road ride has 175 km range).
A typical electric scooter with half-kilowatt-hour battery is advertised to have about 45 km range. However, if you read a fine print it's for a lightweight rider riding at constant 20 km/h speed with no braking, no headwind and perfectly level ground. I actually saw a test report of a half-kilowatt-hour e-scooter that was advertised to have 45 km range, but had only 21 km range at higher speed when braking occasionally and riding occasionally up hills. Also most e-scooters limit the power level to a very low value when only 20% of battery is left. Thus, only 0.8 * 21 km = 16.8 km provides full power. Also typically lithium ion batteries are considered to be at the end of their useful lifetime when 70% of capacity is left. Thus, you can only rely on full power at the end of the battery lifetime to have 0.7 * 16.8 km = 11.76 km range.
Why is the range on an e-bike far better? The reason is based on many differences.
Firstly, a bicycle has very low rolling resistance high pressure narrow large diameter tires. An e-scooter usually has either honeycomb airless tires or maybe low pressure small diameter wide pneumatic tires (in which case it'll be a very major chore to fix a puncture as the electric motor is built into the wheel hub). The e-scooter tires have several times worse rolling resistance than bicycle tires.
Secondly, a bicyclist is riding in a fairly aerodynamic position. In most e-scooters, the rider is standing, maximizing the frontal area.
Thirdly, in electric bikes the cyclist can ride on human power at speeds greater than 25 km/h on flat land. Thus, the electric assist is practically never used on flat land (only momentarily when accelerating to full speed). It's an uphill assist device, and even then only an assist device and not a device that produces 100% of the power. Because of this, the motor is rarely used that increases the range to very high values.
Also consider that the 125 - 175 km range of half-kilowatt-hour e-bikes can be extended if considering acceptable to ride without electric assist. Thus, you can actually rely on the range. If the battery starts to become aged, you can still ride the 125 - 175 km, only the very end of the journey has to be made without electric assist.
In contrast, on a half-kilowatt-hour e-scooter you can only rely on the 11.76 km range on aged battery without using the last 20% that only offers very much reduced power (so the speed will be horribly slow).
Do you want 125 - 175 km range?
Or do you want 11.76 km range?
I'll choose the 125 - 175 km range. I chose an electric bike therefore.
Unfortunately, quality e-bikes are still expensive. Spending too little typically means you have an e-bike that does not have a power sensor on the bottom bracket (so the computer has no idea how hard you are pedaling and the pedaling level doesn't directly affect assist level). Typically on cheap e-bikes the motor is at the front wheel hub. Also on cheap e-bikes the battery is typically very small.
Fortunately, e-bike prices are continuously decreasing. So in few years, a quality mid-drive e-bike with decent half kilowatt hour battery can be bought for far less money than it costs today.
Riding on an e-bike is still considered "exercise" so it has major health benefits. However, by dressing suitably you won't become sweaty like you do on a regular bike if needing to ride uphill.
Also on an e-bike, if choosing a model that has pannier rack or installing an aftermarket pannier rack, you can carry a lot of stuff on the panniers. Far more than you can fit in a backpack, and far heavier stuff. On e-scooter where the rider is standing, typically the only feasible option to carry stuff is a backpack.